blob_id
stringlengths 40
40
| directory_id
stringlengths 40
40
| path
stringlengths 2
327
| content_id
stringlengths 40
40
| detected_licenses
listlengths 0
91
| license_type
stringclasses 2
values | repo_name
stringlengths 5
134
| snapshot_id
stringlengths 40
40
| revision_id
stringlengths 40
40
| branch_name
stringclasses 46
values | visit_date
timestamp[us]date 2016-08-02 22:44:29
2023-09-06 08:39:28
| revision_date
timestamp[us]date 1977-08-08 00:00:00
2023-09-05 12:13:49
| committer_date
timestamp[us]date 1977-08-08 00:00:00
2023-09-05 12:13:49
| github_id
int64 19.4k
671M
⌀ | star_events_count
int64 0
40k
| fork_events_count
int64 0
32.4k
| gha_license_id
stringclasses 14
values | gha_event_created_at
timestamp[us]date 2012-06-21 16:39:19
2023-09-14 21:52:42
⌀ | gha_created_at
timestamp[us]date 2008-05-25 01:21:32
2023-06-28 13:19:12
⌀ | gha_language
stringclasses 60
values | src_encoding
stringclasses 24
values | language
stringclasses 1
value | is_vendor
bool 2
classes | is_generated
bool 2
classes | length_bytes
int64 7
9.18M
| extension
stringclasses 20
values | filename
stringlengths 1
141
| content
stringlengths 7
9.18M
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
066ee8a3941c60fa6c8b5f90a5a701c4287b3f8b
|
b6c263edf819e3cda7264f964bcbbe3a857fd140
|
/R/loc_est_bw.R
|
4ed3aaa5900f9ff7c0b1abc4fdb21e0196b9ff5c
|
[] |
no_license
|
cran/npbr
|
d2da3bfa7c43c255e70651b031de24be85a5d4cf
|
fb273a2137bc9dbd3efe0ce35aa2c6c2fd6fc638
|
refs/heads/master
| 2023-04-10T12:02:30.030340
| 2023-03-22T08:00:05
| 2023-03-22T08:00:05
| 17,697,969
| 1
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 3,183
|
r
|
loc_est_bw.R
|
loc_est_bw <- function(xtab, ytab, x, hini, B = 5, method = "u", fix.seed = FALSE, control = list("tm_limit" = 700)){
# verification
stopifnot(length(xtab)==length(ytab))
# initialiasation
h_loc_min <- max(diff(sort(xtab)))
h_loc_max <- (max(xtab)-min(xtab))/2
h_loc_range <- seq(h_loc_min, h_loc_max, length = 30)
BIterN <- B # number of bootstrap iteration
ghat <- loc_est(xtab, ytab, x, hini, method = method, control = control) # a pilot estimator
gpil <- length(ghat) # a second pilot estimator
h1 <- 1.5*hini
# internal function
ckernel <- function(x, bandwidth){
as.numeric((x/bandwidth<=0.5) & (x/bandwidth>=-0.5))*bandwidth^{-1}
}
for(i in 1:length(x)){
gpil[i] <- sum(ghat*ckernel(x[i]-x, h1), na.rm = T)*unique(diff(x))[1]/(sum(ckernel(x[i]-x, h1), na.rm = T)*unique(diff(x))[1]) # smoothing
}
####################################
# STEA (a) in Hall and Park (2004) #
####################################
rm_ind <- c()
for (i in 1:length(xtab)){
if(ytab[i]>gpil[which.min( abs(xtab[i]-x))]){
rm_ind <- c(rm_ind, i)
}
}
# L_1
xtab_r <- xtab[-rm_ind]
ytab_r <- ytab[-rm_ind]
####################################
# STEA (b) in Hall and Park (2004) #
####################################
# L
xtab_e <- xtab_r
ytab_e <- ytab_r
for(i in 1:length(xtab_r)){
ytab_e <- c(ytab_e, 2*gpil[which.min(abs(xtab_r[i]-x))]-ytab_r[i])
xtab_e <- c(xtab_e, xtab_r[i])
}
######################################
# STEA (c,d) in Hall and Park (2004) #
######################################
m <- as.matrix(dist(cbind(xtab_e, ytab_e)))
sortm <- apply(m, 2, sort)
k <- 10
DZ <- sortm[k+1,]
MSE <- rep(0, length(h_loc_range))
count <- rep(0, length(h_loc_range))
for(iterB in 1:BIterN){
cat("Bootstrap Sample #", iterB, "\n")
if(fix.seed)
set.seed(iterB)
NZ <- rpois(length(xtab_e), 1)
bxtab <- c()
bytab <- c()
for (i in 1:length(xtab_e)){
if(NZ[i]>0){
if(fix.seed)
set.seed(i)
R <- runif(NZ[i], 0, DZ[i])
if(fix.seed)
set.seed(i+1)
theta <- runif(NZ[i], 0, 2*pi)
bxtab <- c(bxtab, xtab_e[i] + R*cos(theta))
bytab <- c(bytab, ytab_e[i] + R*sin(theta))
}
}
rm_ind_b <- c()
for(i in 1:length(bxtab)){
if(bytab[i]>gpil[which.min(abs(bxtab[i]-x))]){
rm_ind_b <- c(rm_ind_b, i)
}
}
bxtab_r <- bxtab[-rm_ind_b]
bytab_r <- bytab[-rm_ind_b]
h_chk_min <- max(diff(sort(bxtab_r)))
if(h_chk_min < max(h_loc_range)){
count <- count + as.numeric(h_loc_range>=h_chk_min)
xb <- x[(x>=min(bxtab_r)) & (x<=max(bxtab_r))]
gpilb <- gpil[(x>=min(bxtab_r)) & (x<=max(bxtab_r))]
for(hi in min(which(h_loc_range>=h_chk_min)):length(h_loc_range)){
est <- loc_est(bxtab_r, bytab_r, xb, h = h_loc_range[hi], method = method, control = control)
MSE[hi] <- MSE[hi] + mean((est-gpilb)^2, na.rm = T)
}
}
}
h_loc_range_r <- h_loc_range[count>0]
hbopt <- h_loc_range_r[which.min(MSE[count>0]*(1/count[count>0]))]
return(hbopt)
}
|
6b4ca402b5a648eada95f0da1c38841626b48f7c
|
7ea08d762a5cfad1ff672199b1431ac9e9449a3f
|
/blog-2023/Blog-4-submissions/charchit/ozone-map.R
|
1ad86591c9640549639cd55cdd34ef1bc09185f1
|
[] |
no_license
|
Stat585-at-ISU/Stat585-at-ISU.github.io
|
54a0d08c274522914aabd2d74e71259d5253b326
|
33d38ef0d67047d83fa8b7a216598e162a5f2400
|
refs/heads/main
| 2023-07-06T14:48:30.349444
| 2023-06-26T20:24:22
| 2023-06-26T20:24:22
| 78,376,998
| 3
| 3
| null | 2017-04-17T04:23:52
| 2017-01-08T23:18:27
|
HTML
|
UTF-8
|
R
| false
| false
| 1,277
|
r
|
ozone-map.R
|
library("ggplot2")
library("maps")
outlines <- as.data.frame(map("world",xlim=-c(113.8, 56.2), ylim=c(-21.2, 36.2), plot=FALSE)[c("x","y")])
map <- c(
geom_path(aes(x=x, y=y, fill=NULL, group=NULL, order=NULL, size=NULL), data = outlines, colour = alpha("grey20", 0.2)),
scale_x_continuous("", limits = c(-114.8, -55.2), breaks=c(-110, -85, -60)),
scale_y_continuous("", limits = c(-22.2, 37.2))
)
ozm <- melt(ozone)
fac <- laply(ozm, is.factor)
ozm[fac] <- llply(ozm[fac], function(x) as.numeric(as.character(x)))
small_mult <- function(df) {
res <- 2 # lat and long measured every 2.5, but need a gap
rexpo <- transform(df,
rozone = rescaler(value, type="range") - 0.5,
rtime = rescaler(time %% 12, type="range") - 0.5,
year = time %/% 12
)
ggplot(rexpo, aes(x = long + res * rtime, y = lat + res * rozone)) + map
}
make_stars <- function(data, time, value) {
data[, c(time, value)] <- lapply(data[, c(time, value)], function(x) {
x <- as.numeric(x)
(x - min(x)) / diff(range(x))
})
ddply(data, .(lat, long), function(df) {
df$x <- df[, value] * cos(df[, time] * 2 * pi + pi / 2)
df$y <- df[, value] * sin(df[, time] * 2 * pi + pi / 2)
df[order(df[, time]), ]
})
}
# df <- stars(owide[, 1:2], owide[, -(1:2)])
|
9463d8e1a5ac4f18cfe3118a7fcd605bf0836fe5
|
cbfa01cb81d4aa3684655ed93b7e179819057083
|
/tests/testthat.R
|
d011b841a13e6a774e2ec3e108abff37a90a99f4
|
[
"MIT"
] |
permissive
|
dirkschumacher/lazyseq
|
b1d00d3603ae845acd495788304af66d8ad30fb1
|
00bde2c17f21a8708c9d6a98666435f7c1f871e1
|
refs/heads/main
| 2023-02-26T14:15:19.264720
| 2021-02-02T20:42:39
| 2021-02-02T20:42:39
| 335,417,878
| 2
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 58
|
r
|
testthat.R
|
library(testthat)
library(lazyseq)
test_check("lazyseq")
|
b5b5f98a971832d5368ecc6b4fc83f5b5b267675
|
b798dcec9242b0656453186201aa2411f65fe28d
|
/man/createProfileMatrix.Rd
|
fd6601f481f257f686c254e98c00cac48ec4e8b5
|
[] |
no_license
|
frosinastojanovska/Bioinformatics
|
481367484796e04fc89ae2675644652355d88729
|
65620c37026336f816d4d981addfc92b15a2d488
|
refs/heads/master
| 2021-01-17T08:42:15.789577
| 2017-03-24T20:08:31
| 2017-03-24T20:08:31
| 83,944,435
| 1
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 537
|
rd
|
createProfileMatrix.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/createProfileMatrix.R
\name{createProfileMatrix}
\alias{createProfileMatrix}
\title{Creates profile matrix for given aligment matrix.}
\usage{
createProfileMatrix(aligmentMatrix)
}
\arguments{
\item{aligmentMatrix}{A matrix indicating the aligment matrix.}
}
\value{
Profile matrix
}
\description{
Creates profile matrix for given aligment matrix.
}
\examples{
createProfileMatrix(matrix(c("A","C","A","C","A","C","U","A","A","C","U","A"), nrow=4, ncol=3))
}
|
6cf98ade83431b0e5e5e728adfabcde688d9d20d
|
24975c66d61805ffd50147890b9fc34769f18324
|
/Notes_scripts_bank/ex3_binomial-uniform_multiple_control_JAGS.R
|
43d5bfd580af59871877c1d8ad62400cf2a66aff
|
[] |
no_license
|
npetraco/MATFOS705
|
53081de4e38a1aae8e0d67bf093a1bcd6b9f0258
|
dc54407b7b13ebf8315282cbaf0c9b742212884d
|
refs/heads/master
| 2023-05-28T03:08:46.839815
| 2023-05-16T14:47:15
| 2023-05-16T14:47:15
| 121,066,670
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,352
|
r
|
ex3_binomial-uniform_multiple_control_JAGS.R
|
library(bayesutils)
# Data
n <- rep(50, 5)
participant.errs <- c(0, 0, 20, 0, 1)
s <- participant.errs
dat <- list(
"N" = length(n),
"s" = s,
"n" = n,
"a_hyp" = 0,
"b_hyp" = 1
)
inits <- function (){
list(ppi=runif(length(n)))
}
#Run the model:
fit <- jags(data=dat,
inits=inits,
parameters.to.save = c("ppi","mean.ppi"),
#n.iter=10,
n.iter=20000, n.burnin = 500, n.thin = 10,
n.chains=4,
model.file = system.file("jags/binomial-uniform_multiple.bug.R", package = "bayesutils"))
fit
# Examine chains trace and autocorrelation:
#params.chains <- extract.params(fit, by.chainQ = T)
#mcmc_trace(params.chains, regex_pars = c("ppi"))
#autocorrelation.plots(params.chains, pars = c("ppi"))
# Examine posterior
params.mat <- extract.params(fit, as.matrixQ = T)
mcmc_areas(params.mat, regex_pars = c("ppi"), prob = 0.95)
dim(params.mat)
colnames(params.mat)
ppi <- params.mat[,2:6]
mean.ppi <- params.mat[,1]
parameter.intervals(ppi[,1], prob=0.95, plotQ = F)
parameter.intervals(ppi[,2], prob=0.95, plotQ = F)
parameter.intervals(ppi[,3], prob=0.95, plotQ = F)
parameter.intervals(ppi[,4], prob=0.95, plotQ = F)
parameter.intervals(ppi[,5], prob=0.95, plotQ = F)
parameter.intervals(mean.ppi, prob=0.95, plotQ = F)
|
57da56bac26b30b5c61af0cfa6ff79342b64cb42
|
24d5d85c21f8843f7c9c61e00222522fde104e7d
|
/percentlabelv2.R
|
8322c2d37624286fabf594b137a392d3b152fb0b
|
[] |
no_license
|
cassierole/ms_percentlabel
|
deab302aee17274f2cfd407aae2e7429609a13ca
|
25cd2408668979e56f4765e3897cc3b381823a46
|
refs/heads/master
| 2021-01-10T13:30:26.909309
| 2016-04-11T23:19:49
| 2016-04-11T23:19:49
| 52,817,206
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 3,023
|
r
|
percentlabelv2.R
|
# This function accepts as argument a csv file (ex."file.csv") containing raw quantitative
# mass spec data and produces a file in the working directory containing the percentage of
# C13-labeled phospholipid for each individual species.
# Experiments carried out on Acinetobacter baumannii fed 2-C13 acetate
percentlabelv2 <- function(datafile){
dat <- read.csv(datafile, check.names = FALSE) #import raw mass spec data
frame <- data.frame() #create an empty data frame for binding
cname <-names(dat) #store the column names from mass spec data in vector cnames
for (i in 1:(nrow(dat))){
name <- dat[i,grep("name",cname,ignore.case=TRUE)]
#calculate the percentage labeled for each species and save as variables
dioPG <- getvalues(i,"795","773",dat,cname)
paloPG <- getvalues(i,"768","747",dat,cname)
dipPG <- getvalues(i,"739","719",dat,cname)
dioPE <- getvalues(i,"764","742",dat,cname)
paloPE <- getvalues(i,"737","716",dat,cname)
dipPE <- getvalues(i,"708","688",dat,cname)
card <- getvalues(i,"711","701",dat,cname)
#create a new data frame containing percentages, then bind this to existing data frame
newframe <- data.frame("Sample_Name"=name,"PG C18:1/18:1"=dioPG,"PG C18:1/16:0"=paloPG,"PG C16:0/16:0"=dipPG,"PE C18:1/18:1"=dioPE,"PE C18:1/16:0"=paloPE,"PE C16:0/16:0"=dipPE,"Cardiolipin"=card, check.names=FALSE)
frame <- rbind(frame,newframe)
}
#creates a file in working directory containing percentages:
filename <- paste("percent_labeled",datafile,sep="_")
write.csv(frame,file=filename)
print("Calculation Complete")
#frame
}
#a function to calculate percentages
calcper <- function(lab, unlab){
perlab <- 100 * lab / (lab + unlab)
}
#getvalues accepts as arguments row of database, m/z of parent ions for labeled and unlabeled, and database
#returns percentage of labeled species out of total for that species.
getvalues <-function(row,label,unlabel,data,cnames){
#store as a vectors the column positions of labeled and unlabeled species
loclab <- grep(label, cnames, ignore.case=TRUE)
locunl <- grep(unlabel, cnames, ignore.case=TRUE)
vallab <- NA
valunl <- NA
#If there is more than one transition corresponding to a given parent ion, sum the quantitation values
if ((length(loclab)>0)&&(length(locunl)>0)){
for (i in 1:length(loclab)){
vallab <- sum(vallab, data[row,loclab[i]],na.rm=TRUE)
}
for (i in 1:length(locunl)){
valunl <- sum(valunl, data[row,locunl[i]],na.rm=TRUE)
}
}
calcper(vallab,valunl)
}
|
ac065990bb0a476637a982b3d38e77c84571e1fc
|
799f724f939763c26c4c94497b8632bad380e8f3
|
/man/as.tokens.Rd
|
c1ad9479f2599038451fba137887d45206cbc3fd
|
[] |
no_license
|
chmue/quanteda
|
89133a7196b1617f599e5bba57fe1f6e59b5c579
|
aed5cce6778150be790b66c031ac8a40431ec712
|
refs/heads/master
| 2020-12-01T02:59:48.346832
| 2017-11-22T18:35:37
| 2017-11-22T18:35:37
| 50,363,453
| 2
| 0
| null | 2016-01-25T16:18:50
| 2016-01-25T16:18:50
| null |
UTF-8
|
R
| false
| true
| 3,635
|
rd
|
as.tokens.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/tokens.R
\name{as.tokens}
\alias{as.tokens}
\alias{as.tokens.list}
\alias{as.tokens.spacyr_parsed}
\alias{as.list.tokens}
\alias{unlist.tokens}
\alias{as.character.tokens}
\alias{is.tokens}
\alias{+.tokens}
\alias{c.tokens}
\title{coercion, checking, and combining functions for tokens objects}
\usage{
as.tokens(x, concatenator = "_", ...)
\method{as.tokens}{list}(x, concatenator = "_", ...)
\method{as.tokens}{spacyr_parsed}(x, concatenator = "/",
include_pos = c("none", "pos", "tag"), use_lemma = FALSE, ...)
\method{as.list}{tokens}(x, ...)
\method{unlist}{tokens}(x, recursive = FALSE, use.names = TRUE)
\method{as.character}{tokens}(x, use.names = FALSE, ...)
is.tokens(x)
\method{+}{tokens}(t1, t2)
\method{c}{tokens}(...)
}
\arguments{
\item{x}{object to be coerced or checked}
\item{concatenator}{character between multi-word expressions, default is the
underscore character. See Details.}
\item{...}{additional arguments used by specific methods. For
\link{c.tokens}, these are the \link{tokens} objects to be concatenated.}
\item{include_pos}{character; whether and which part-of-speech tag to use:
\code{"none"} do not use any part of speech indicator, \code{"pos"} use the
\code{pos} variable, \code{"tag"} use the \code{tag} variable. The POS
will be added to the token after \code{"concatenator"}.}
\item{use_lemma}{logical; if \code{TRUE}, use the lemma rather than the raw
token}
\item{recursive}{a required argument for \link{unlist} but inapplicable to
\link{tokens} objects}
\item{use.names}{logical; preserve names if \code{TRUE}. For
\code{as.character} and \code{unlist} only.}
\item{t1}{tokens one to be added}
\item{t2}{tokens two to be added}
}
\value{
\code{as.tokens} returns a quanteda \link{tokens} object.
\code{as.list} returns a simple list of characters from a
\link{tokens} object.
\code{unlist} returns a simple vector of characters from a
\link{tokens} object.
\code{as.character} returns a character vector from a
\link{tokens} object.
\code{is.tokens} returns \code{TRUE} if the object is of class
tokens, \code{FALSE} otherwise.
\code{c(...)} and \code{+} return a tokens object whose documents
have been added as a single sequence of documents.
}
\description{
Coercion functions to and from \link{tokens} objects, checks for whether an
object is a \link{tokens} object, and functions to combine \link{tokens}
objects.
}
\details{
The \code{concatenator} is used to automatically generate dictionary
values for multi-word expressions in \code{\link{tokens_lookup}} and
\code{\link{dfm_lookup}}. The underscore character is commonly used to join
elements of multi-word expressions (e.g. "piece_of_cake", "New_York"), but
other characters (e.g. whitespace " " or a hyphen "-") can also be used.
In those cases, users have to tell the system what is the concatenator in
your tokens so that the conversion knows to treat this character as the
inter-word delimiter, when reading in the elements that will become the
tokens.
}
\examples{
# create tokens object from list of characters with custom concatenator
dict <- dictionary(list(country = "United States",
sea = c("Atlantic Ocean", "Pacific Ocean")))
lis <- list(c("The", "United-States", "has", "the", "Atlantic-Ocean",
"and", "the", "Pacific-Ocean", "."))
toks <- as.tokens(lis, concatenator = "-")
tokens_lookup(toks, dict)
# combining tokens
toks1 <- tokens(c(doc1 = "a b c d e", doc2 = "f g h"))
toks2 <- tokens(c(doc3 = "1 2 3"))
toks1 + toks2
c(toks1, toks2)
}
|
c4e710346981ed9c21da1c4701381fdd96659e9d
|
f97f007dc8fab3d266fa9426f1dc612ba23754e7
|
/man/high_value_terms.Rd
|
221a9c828c2be02401b00a5f0723d75f088d3375
|
[] |
no_license
|
anpatton/overdoseR
|
6e5c7b8f66d9664ee455379ab23392918d138530
|
e84e15a027b18fff9530950aaef23fcc350e2455
|
refs/heads/master
| 2022-11-16T11:09:32.289312
| 2020-07-07T16:38:51
| 2020-07-07T16:38:51
| 276,998,525
| 2
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 686
|
rd
|
high_value_terms.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/data.R
\docType{data}
\name{high_value_terms}
\alias{high_value_terms}
\title{High Value Terms}
\format{
A data frame with 189 rows and 4 variables:
\describe{
\item{token}{actual string of high value term}
\item{freqYES}{Frequency of term in development set for opioid related events}
\item{freqNO}{Frequency of term in development set for non-opioid related events}
\item{type}{word, bigram, or trigram}
}
}
\usage{
high_value_terms
}
\description{
Dataset containing the raw words, bigrams, and trigrams that
were determined to be statistically significant in prior research.
}
\keyword{datasets}
|
8b037e9259022527bcc68a473bceedb6f854b9c6
|
7f3d289e75c1faf4a44d50621d969dc70658259b
|
/2-StackFiles.R
|
e7dddddc2ec26fee5809d6330040f4c48275d23c
|
[] |
no_license
|
alessiobocco/IFPRI_Ethiopia_Drought_2016
|
89ccaf8bbbdbe9cee57925ed4e41c8bded0f99f5
|
5f08c729c1e567adc9c629162d10b90e7ec09470
|
refs/heads/master
| 2020-05-27T21:09:39.770253
| 2017-02-01T14:25:57
| 2017-02-01T14:25:57
| 83,603,162
| 1
| 0
| null | 2017-03-01T21:23:58
| 2017-03-01T21:23:58
| null |
UTF-8
|
R
| false
| false
| 14,948
|
r
|
2-StackFiles.R
|
# This script takes outputs from DownloadMODISFTP_Rcurl.R and stacks them
# Run the following in bash before starting R
if [ -e $HOME/.Renviron ]; then cp $HOME/.Renviron $HOME/.Renviron.bkp; fi
if [ ! -d $HOME/.Rtmp ] ; then mkdir $HOME/.Rtmp; fi
echo "TMP='$HOME/.Rtmp'" > $HOME/.Renviron
module load proj.4/4.8.0
module load gdal/gcc/1.11
module load R
module load gcc/4.9.0
R
rm(list=ls())
#source('R:\\Mann Research\\IFPRI_Ethiopia_Drought_2016\\IFPRI_Ethiopia_Drought_Code\\ModisDownload.R')
source('/groups/manngroup/IFPRI_Ethiopia_Dought_2016/IFPRI_Ethiopia_Drought_2016/SummaryFunctions.R')
library(RCurl)
library(raster)
library(MODISTools)
library(rgdal)
library(sp)
library(maptools)
#library(rts)
library(gdalUtils)
library(foreach)
library(doParallel)
library(compiler)
library(ggplot2)
#cl <- makeCluster(32)
#registerDoParallel(cl)
# Compile Functions ---------------------------------------------------------------
functions_in = lsf.str()
lapply(1:length(functions_in), function(x){cmpfun(get(functions_in[[x]]))}) # byte code compile all functions http://adv-r.had.co.nz/Profil$
# Set up parameters -------------------------------------------------------
# give path to Modis Reproduction Tool
MRT = 'H:/Projects/MRT/bin'
# get list of all available modis products
#GetProducts()
# Product Filters
products = c('MYD13Q1') #EVI c('MYD13Q1','MOD13Q1') , land cover = 'MCD12Q1' for 250m and landcover ='MCD12Q2'
location = c(9.145000, 40.489673) # Lat Lon of a location of interest within your tiles listed above #India c(-31.467934,-57.101319) #
tiles = c('h21v07','h22v07','h21v08','h22v08') # India example c('h13v12')
dates = c('2011-01-01','2016-03-30') # example c('year-month-day',year-month-day') c('2002-07-04','2016-02-02')
ftp = 'ftp://ladsweb.nascom.nasa.gov/allData/6/' # allData/6/ for evi, /51/ for landcover
# allData/51/ for landcover DOESn't WORK jUST PULL FROM FTP
strptime(gsub("^.*A([0-9]+).*$", "\\1",GetDates(location[1], location[2],products[1])),'%Y%j') # get list of all available dates for products[1]
# out_dir = 'R:\\Mann_Research\\IFPRI_Ethiopia_Drought_2016\\Data\\VegetationIndex'
# setwd(out_dir)
# Stack Raw data -----------------------------------------------------
registerDoParallel(5)
setwd('/groups/manngroup/IFPRI_Ethiopia_Dought_2016/Data/VegetationIndex/') # folder where EVI .tifs are
# create data stack for each variable and tile
foreach(product = c('composite_day_of_the_year',
'EVI','NDVI','pixel_reliability')) %dopar% {
for( tile_2_process in tiles){
# Set up data
print('stacking')
flist = list.files(".",glob2rx(paste('*',tile_2_process,'.250m_16_days_',product,'.tif$',sep='')),
full.names = TRUE)
flist_dates = gsub("^.*_([0-9]{7})_.*$", "\\1",flist,perl = T) # Strip dates
flist = flist[order(flist_dates)] # file list in order
flist_dates = flist_dates[order(flist_dates)] # file_dates list in order
# stack data and save
stacked = stack(flist)
names(stacked) = flist_dates
# assign projection
crs(stacked) ='+proj=sinu +a=6371007.181 +b=6371007.181 +units=m'
# save
assign(paste(product,'stack',tile_2_process,sep='_'),stacked)
dir.create(file.path('../Data Stacks/Raw Stacks/'), showWarnings=F,recursive=T) # create stack directory if doesnt exist
save( list=paste(product,'stack',tile_2_process,sep='_') ,
file = paste('../Data Stacks/Raw Stacks/',product,'_stack_',tile_2_process,'.RData',sep='') )
}}
# Limit stacks to common dates -------------------------------------------
setwd('/groups/manngroup/IFPRI_Ethiopia_Dought_2016/Data/')
# load data stacks from both directories
stack_types_2_load =c('composite_day_of_the_year','EVI','NDVI','pixel_reliability')
dir1 = list.files('./Data Stacks/Raw Stacks/','.RData',full.names=T)
lapply(dir1, load,.GlobalEnv)
# limit stacks to common elements
for(product in stack_types_2_load ){
for( tile in tiles){
# find dates that exist in all datasets for current tile
all_dates = lapply(paste(stack_types_2_load,'stack',tile,sep='_'),function(x){names(get(x))})
# restrict to common dates
common_dates = Reduce(intersect, all_dates)
# subset stacks for common dates
assign(paste(product,'_stack_',tile,sep=''),subset( get(paste(product,'_stack_',tile,sep='')),
common_dates, drop=F) )
print('raster depth all equal')
print( all.equal(common_dates,names(get(paste(product,'_stack_',tile,sep='')))) )
print(dim(get(paste(product,'_stack_',tile,sep='')))[3])
}}
# stack smoother -----------------------------------------------------
# this stack is used for land cover classification only (bc classifier can't have NA values)
rm(list=ls()[grep('stack',ls())]) # running into memory issues clear stacks load one by one
setwd('/groups/manngroup/IFPRI_Ethiopia_Dought_2016/Data/Data Stacks/Raw Stacks/') # don't load smoothed...
# load data stacks from both directories
dir1 = list.files('.','.RData',full.names=T)
lapply(dir1, load,.GlobalEnv)
for( i in ls(pattern = "NDVI_stack*")){
print('##############################################################')
dir.create(file.path(getwd(), i), showWarnings = FALSE)
print(paste('Starting processing of:',i))
stack_in = get(i)
stack_name = i
dates = as.numeric(gsub("^.*X([0-9]{7}).*$", "\\1",names(stack_in),perl = T)) # Strip dates
pred_dates = dates
spline_spar=0.4 # 0.4 for RF
workers = 20
out_dir = '/groups/manngroup/IFPRI_Ethiopia_Dought_2016/Data/Data Stacks/Smoothed/'
stack_smoother(stack_in,dates,pred_dates,spline_spar,workers,stack_name,out_dir)
}
for( i in ls(pattern = "EVI_stack*")){
print('##############################################################')
dir.create(file.path(getwd(), i), showWarnings = FALSE)
print(paste('Starting processing of:',i))
stack_in = get(i)
stack_name = i
dates = as.numeric(gsub("^.*X([0-9]{7}).*$", "\\1",names(stack_in),perl = T)) # Strip dates
pred_dates = dates
spline_spar=0.4 # 0.4 for RF
workers = 20
out_dir = '/groups/manngroup/IFPRI_Ethiopia_Dought_2016/Data/Data Stacks/Smoothed/'
stack_smoother(stack_in,dates,pred_dates,spline_spar,workers,stack_name,out_dir)
}
# Restack Smoothed Files ----------------------------------------------------
setwd('/groups/manngroup/IFPRI_Ethiopia_Dought_2016/Data/Data Stacks/Smoothed/Tifs/') # folder where EVI .tifs are
# create data stack for each variable and tile
# load data stacks from both directories
dir1 = list.files('.','.RData',full.names=T)
lapply(dir1, load,.GlobalEnv)
foreach(product = c('NDVI','EVI')) %do% {
for( tile_2_process in tiles){
print(paste('processing',product,tile_2_process,sep=' '))
# Set up data
flist = list.files(".",glob2rx(paste(product,'_',tile_2_process,'*','.tif$',sep='')),
full.names = TRUE)
flist_dates = gsub("^.*_X([0-9]{7}).*$", "\\1",flist,perl = T) # Strip dates
flist = flist[order(flist_dates)] # file list in order
flist_dates = flist_dates[order(flist_dates)] # file_dates list in order
# stack data and save
stacked = stack(flist)
names(stacked) = flist_dates
assign(paste(product,'stack',tile_2_process,'smooth',sep='_'),stacked)
save( list=paste(product,'stack',tile_2_process,'smooth',sep='_') ,
file = paste('/groups/manngroup/IFPRI_Ethiopia_Dought_2016/Data/Data Stacks/Smoothed/',
product,'_stack_',tile_2_process,'_smooth','.RData',sep='') )
}}
# Remove low quality,CLEAN, & assign projection FROM RAW, THEN STACK ------------------------------------------------
# load data in previous section and run common dates
rm(list=ls()[grep('stack',ls())]) # running into memory issues clear stacks load one by one
setwd('/groups/manngroup/IFPRI_Ethiopia_Dought_2016/Data/Data Stacks/Raw Stacks/') # don't load smoothed...
# load data stacks from both directories
dir1 = list.files('.','.RData',full.names=T)
lapply(dir1, load,.GlobalEnv)
# set up directories and names
setwd('/groups/manngroup/IFPRI_Ethiopia_Dought_2016/Data//Data Stacks')
reliability_prefix = 'pixel_reliability'
dir.create(file.path('./WO Clouds/Tifs'), showWarnings=F,recursive=T)
dir.create(file.path('./WO Clouds Clean/tifs'), showWarnings=F,recursive=T)
dir.create(file.path('/lustre/groups/manngroup/WO Clouds Clean/Tifs'), showWarnings=F,recursive=T) # folder on high speed ssd drive
registerDoParallel(25)
# setup a dataframe with valid ranges and scale factors
valid = data.frame(stack='NDVI', fill= -3000,validL=-2000,validU=10000,
scale=0.0001,stringsAsFactors=F)
valid = rbind(valid,c('EVI',-3000,-2000,10000,0.0001))
valid
for(product in c('EVI','NDVI')){ #'EVI','NDVI'
for( tile in tiles){
print(paste('Working on',product,tile))
# load quality flag
reliability_stackvalues = get(paste(reliability_prefix,'_stack_',tile,sep=''))
# remove clouds from produt
data_stackvalues = get(paste(product,'_stack_',tile,sep=''))
valid_values = valid[grep(product,valid$stack),]
ScaleClean = function(x,y){
x[x==as.numeric(valid_values$fill)]=NA
x[x < as.numeric(valid_values$validL)]=NA
x[x > as.numeric(valid_values$validU)]=NA
#x = x * as.numeric(valid_values$scale)
x[ y<0 | y>1 ] = NA # remove very low quality
x}
# process and write to lustre
foreach(i=(1:dim(data_stackvalues)[3]), .inorder=F) %dopar% {
print(i)
data_stackvalues[[i]] = ScaleClean(data_stackvalues[[i]],reliability_stackvalues[[i]])
writeRaster(data_stackvalues[[i]],paste('/lustre/groups/manngroup/WO Clouds Clean/Tifs/',product,'_',tile,
'_',names(data_stackvalues[[i]]),'.tif',sep=''),overwrite=T)
}
# Copy files back from lustre and delete lustre
flist = list.files("/lustre/groups/manngroup/WO Clouds Clean/Tifs/",
glob2rx(paste(product,'_',tile,'*','.tif$',sep='')),full.names = T)
fname = list.files("/lustre/groups/manngroup/WO Clouds Clean/Tifs/",
glob2rx(paste(product,'_',tile,'*','.tif$',sep='')),full.names = F)
file.copy(from=flist, to=paste("./WO Clouds Clean/tifs",fname,sep='/'),
overwrite = T, recursive = F, copy.mode = T)
file.remove(flist)
# Restack outputs
print(paste('Restacking',product,tile,sep=' '))
# Set up data
flist = list.files("./WO Clouds Clean/tifs/",glob2rx(paste(product,'_',tile,'*','.tif$',sep='')),full.names = T)
flist_dates = gsub("^.*_X([0-9]{7}).*$", "\\1",flist,perl = T) # Strip dates
flist = flist[order(flist_dates)] # file list in order
flist_dates = flist_dates[order(flist_dates)] # file_dates list in order
# stack data and save
stacked = stack(flist)
names(stacked) = flist_dates
assign(paste(product,'stack',tile,'wo_clouds_clean',sep='_'),stacked)
save( list=paste(product,'stack',tile,'wo_clouds_clean',sep='_') ,
file = paste('./WO Clouds Clean/',product,'_stack_',
tile,'_wo_clouds_clean','.RData',sep='') )
}}
# Limit to crop signal ----------------------------------------------------
# Class Codes:
# 1 agforest 2 arid 3 dryag 4 forest 5 semiarid 6 shrub 7 water 8 wetag 9 wetforest
# load data in previous section and run common dates
rm(list=ls()[grep('stack',ls())]) # running into memory issues clear stacks load one by one
setwd('/groups/manngroup/IFPRI_Ethiopia_Dought_2016/Data/Data Stacks/WO Clouds Clean/') # don't load smoothed...
dir.create(file.path('../WO Clouds Clean LC/tifs/'), showWarnings=F,recursive=T) # create dir for tifs
dir.create(file.path('/lustre/groups/manngroup/WO Clouds Clean LC/Tifs'), showWarnings=F,recursive=T) # folder on high speed
# load data stacks from both directories
dir1 = list.files('.','.RData',full.names=T)
lapply(dir1, load,.GlobalEnv)
# set up directories and names
setwd('/groups/manngroup/IFPRI_Ethiopia_Dought_2016/Data//Data Stacks')
landcover_prefix = 'smooth_lc_svm_mn.tif'
landcover_path = '../LandUseClassifications/'
registerDoParallel(25)
for(product in c('EVI','NDVI')){ #'EVI','NDVI'
for( tile in tiles){
print(paste('Working on',product,tile))
# load quality flag
lc_stackvalues = raster(paste(landcover_path,'NDVI','_stack_',tile,'_',landcover_prefix,sep=''))
# remove clouds from produt
data_stackvalues = get(paste(product,'_stack_',tile,'_wo_clouds_clean',sep=''))
foreach(i=(1:dim(data_stackvalues)[3]), .inorder=F) %dopar% {
print(i)
data_stackvalues[[i]][lc_stackvalues[[i]]==2|lc_stackvalues[[i]]==7|
lc_stackvalues[[i]]==9]=NA
writeRaster(data_stackvalues[[i]],paste('/lustre/groups/manngroup/WO Clouds Clean LC/Tifs/'
,product,'_',tile,'_',names(data_stackvalues[[i]]),
'_Clean_LC','.tif',sep=''),overwrite=T)
}
# Copy files back from lustre and delete lustre
flist = list.files("/lustre/groups/manngroup/WO Clouds Clean LC/Tifs/",
glob2rx(paste(product,'_',tile,'*','.tif$',sep='')),full.names = T)
fname = list.files("/lustre/groups/manngroup/WO Clouds Clean LC/Tifs/",
glob2rx(paste(product,'_',tile,'*','.tif$',sep='')),full.names = F)
file.copy(from=flist, to=paste("./WO Clouds Clean LC/tifs",fname,sep='/'),
overwrite = T, recursive = F, copy.mode = T)
file.remove(flist)
print(paste('Restacking',product,tile,sep=' '))
# Set up data
flist = list.files("./WO Clouds Clean LC/tifs/",glob2rx(paste(product,'_',tile,'*','.tif$',sep='')),full.names = T)
flist_dates = gsub("^.*_X([0-9]{7}).*$", "\\1",flist,perl = T) # Strip dates
flist = flist[order(flist_dates)] # file list in order
flist_dates = flist_dates[order(flist_dates)] # file_dates list in order
# stack data and save
stacked = stack(flist)
names(stacked) = flist_dates
assign(paste(product,'stack',tile,'WO_Clouds_Clean_LC',sep='_'),stacked)
save( list=paste(product,'stack',tile,'WO_Clouds_Clean_LC',sep='_') ,
file = paste('./WO_Clouds_Clean_LC/',product,'_stack_',
tile,'_WO_Clouds_Clean_LC','.RData',sep='') )
}}
|
faa1d0ae0f21926af9977941fa71f7a0658f2d76
|
77109777ecf9aa8b467b7225589daecd9f309148
|
/Projeto2.R
|
e741f0e766ca95947928273632630c5c6c22b153
|
[] |
no_license
|
edufrigini/prevendo_estoque
|
5bf889d44cb1e60438433e9853d75c26e4f835cf
|
bf6477529175c58edac3f1a9bae6efc8b1443161
|
refs/heads/master
| 2020-05-25T00:00:19.835408
| 2019-05-19T20:46:54
| 2019-05-19T20:46:54
| 187,526,262
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 7,906
|
r
|
Projeto2.R
|
# DSA - DATA SCIENCE ACADEMY
# FORMACAO CIENTISTA DE DADOS
# LIGUAGEM R COM AZURE MACHINE LEARNING
#
# PROJETO 2, Prevendo Demanda de Estoque com Base em Vendas
# ALUNO: EDUARDO FRIGINI DE JESUS
#
# Goal: Maximize sales and minimize returns of bakery goods
# Data fields
# Semana — Week number (From Thursday to Wednesday)
# Agencia_ID — Sales Depot ID
# Canal_ID — Sales Channel ID
# Ruta_SAK — Route ID (Several routes = Sales Depot)
# Cliente_ID — Client ID
# NombreCliente — Client name
# Producto_ID — Product ID
# NombreProducto — Product Name
# Venta_uni_hoy — Sales unit this week (integer)
# Venta_hoy — Sales this week (unit: pesos)
# Dev_uni_proxima — Returns unit next week (integer)
# Dev_proxima — Returns next week (unit: pesos)
# Demanda_uni_equil — Adjusted Demand (integer) (This is the target you will predict)
setwd("C:/FCD/BigDataRAzure/Projeto2")
getwd()
install.packages("gmodels")
install.packages("psych")
install.packages("rmarkdown")
library(rmarkdown)
# carregando as bibliotecas, se nao estiver instalada, instalar install.packages("nome do pacote")
library(data.table) # para usar a fread
library("gmodels") # para usar o CrossTable
library(psych) # para usar o pairs.panels
library(lattice) # graficos de correlacao
require(ggplot2)
library(randomForest)
library(DMwR)
library(dplyr)
library(tidyr)
library("ROCR")
library(caret)
library(lattice)
library(corrplot)
library(corrgram)
## Carregando os dados na memoria
# Usando o arquivo train.csv para treinar o modelo para producao
dados_originais <- fread("train.csv", sep = ",", header = TRUE, stringsAsFactors = TRUE)
dados <- dados_originais[sample(1:nrow(dados_originais), 1000, replace = F)]
head(dados)
str(dados)
View(dados)
## Tratando os dados
## Convertendo as variáveis para o tipo fator (categórica)
to.factors <- function(df, variables){
for (variable in variables){
df[[variable]] <- as.factor(df[[variable]])
}
return(df)
}
# Variáveis do tipo fator
# nao converti as outras pq eram mts categorias e a floresta randomica nao processa mts categorias
colunas_F <- c("Semana", "Canal_ID")
dados <- to.factors(df = dados, variables = colunas_F)
# corrigir caso hajam dados NA
dados <- na.omit(dados)
str(dados)
head(dados)
## Analise exploratoria dos dados
summary(dados$Demanda_uni_equil)
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.000 2.000 3.000 7.018 6.000 230.000
mean(dados$Demanda_uni_equil) # 7.018
median(dados$Demanda_uni_equil) # 3
quantile(dados$Demanda_uni_equil) # 0% 25% 50% 75% 100%
# 0 2 3 6 230
quantile(dados$Demanda_uni_equil, probs = c(0.01, 0.95)) # 1% = 0.00 e 95% = 23.05
quantile(dados$Demanda_uni_equil, seq(from = 0, to = 1, by = 0.10))
IQR(dados$Demanda_uni_equil) # diferenca entre Q3 e Q1 = 4
range(dados$Demanda_uni_equil) # 230
diff(range(dados$Demanda_uni_equil))
sd(dados$Demanda_uni_equil) # 13.74
var(dados$Demanda_uni_equil) # 188.9666
# A variavel alvo esta com muitos outlyers, dai vou restringir ao valores menores que 0.95 quartil
quartil_90 <- quantile(dados$Demanda_uni_equil, probs = 0.90)
class(quartil_90)
quartil_90[[1]]
dados_sem_ouliers <- dados[Demanda_uni_equil<=quartil_90[[1]],]
plot(dados_sem_ouliers$Demanda_uni_equil)
sd(dados_sem_ouliers$Demanda_uni_equil) # 4.35
var(dados_sem_ouliers$Demanda_uni_equil) # 19
## OS DADOS DO TARGET NAO SEGUEM UMA DISTRIBUICAO NORMAL
## Explorando os dados graficamente
plot(dados_sem_ouliers$Semana)
hist(dados_sem_ouliers$Ruta_SAK)
plot(dados_sem_ouliers$Demanda_uni_equil) # target e esta com outliers
hist(dados_sem_ouliers$Demanda_uni_equil) # dados concentrados no zero
boxplot(dados_sem_ouliers$Demanda_uni_equil)
plot(dados_sem_ouliers$Agencia_ID)
plot(dados_sem_ouliers$Canal_ID) # Predominancia do canal 1
# Explorando os dados
# variaveis numericas
cols <- c("Venta_uni_hoy", "Venta_hoy", "Dev_uni_proxima", "Dev_proxima", "Demanda_uni_equil")
cor(dados[, cols, with=FALSE])
pairs.panels(dados[, cols, with=FALSE])
# Correlacao forte com Venta_Roy e Venta_Uni_Roy
#### OBSERVACOES
#### Venta_uni_hoy e Venta_hoy sao colineares
#### Dev_uni_proxima e Dev Proxima sao colineares
######################################################################
## Apenas para confirmar a correlacao com Venta_hoy e Venta_uni_hoy ##
######################################################################
# Vetor com os métodos de correlação
metodos <- c("pearson", "spearman")
# Aplicando os métodos de correlação com a função cor()
cors <- lapply(metodos, function(method)
(cor(dados[, cols, with=FALSE], method = method)))
head(cors)
# Preparando o plot
plot.cors <- function(x, labs){
diag(x) <- 0.0
plot( levelplot(x,
main = paste("Plot de Correlação usando Método", labs),
scales = list(x = list(rot = 90), cex = 1.0)) )
}
# Mapa de Correlação
Map(plot.cors, cors, metodos)
#########################################################
## Sem necessidade desse grafico, apenas para confirmar
#########################################################
str(dados_sem_ouliers)
# Demanda x potenciais variáveis preditoras
labels <- list("Boxplots - Demanda por Semana",
"Boxplots - Demanda por Agencia",
"Boxplots - Demanda por Canal",
"Boxplots - Demanda por Rota",
"Boxplots - Demanda Cliente")
xAxis <- list("Semana", "Agencia_ID", "Canal_ID", "Ruta_SAK", "Cliente_ID")
# Função para criar os boxplots
plot.boxes <- function(X, label){
ggplot(dados, aes_string(x = X, y = "Demanda_uni_equil", group = X)) +
geom_boxplot( ) +
ggtitle(label) +
theme(text = element_text(size = 18))
}
Map(plot.boxes, xAxis, labels)
str(dados_sem_ouliers)
# Avalidando a importância de todas as variaveis
modelo <- randomForest(Demanda_uni_equil ~ . ,
data = dados_sem_ouliers,
ntree = 100,
nodesize = 10,
importance = TRUE)
# Plotando as variáveis por grau de importância
varImpPlot(modelo)
# Correlacao forte com Venta_Roy e Venta_Uni_Roy
#### OBSERVACOES
#### Venta_uni_hoy e Venta_hoy sao colineares
#### Dev_uni_proxima e Dev_Proxima sao colineares
# Removendo variáveis colineares
modelo <- randomForest(Demanda_uni_equil ~ . - Venta_hoy
- Dev_proxima,
data = dados_sem_ouliers,
ntree = 100,
nodesize = 10,
importance = TRUE)
varImpPlot(modelo)
modelo$importance
# Gravando o resultado
df_saida <- dados_sem_ouliers[, c("Demanda_uni_equil", (modelo$importance))]
df_saida
# Removendo as variaveis menos importantes ou colineares
dados_ok <- dados_sem_ouliers
dados_ok$Dev_proxima <- NULL
dados_ok$Venta_hoy <- NULL
dados_ok$Cliente_ID <- NULL
dados_ok$Agencia_ID <- NULL
str(dados_ok)
# Gerando dados de treino e de teste
sample <- sample.int(n = nrow(dados_ok), size = floor(.7*nrow(dados_sem_ouliers)), replace = F)
treino <- dados_ok[sample, ]
teste <- dados_ok[-sample, ]
# Verificando o numero de linhas
nrow(treino)
nrow(teste)
# Treinando o modelo linear (usando os dados de treino)
modelo_lm <- lm(Demanda_uni_equil ~ ., data = treino)
modelo_lm
# Prevendo demanda de produtos
previsao1 <- predict(modelo_lm)
View(previsao1)
plot(treino$Demanda_uni_equil, previsao1)
previsao2 <- predict(modelo_lm, teste)
View(previsao2)
plot(teste$Demanda_uni_equil, previsao2)
# Avaliando a Performance do Modelo
summary(modelo_lm)
# Adjusted R-squared: 0.9923
|
67225c32124b38f49e03a442d7fbfd4a6dc3ffd1
|
b29688e753d4ae51662783aa23198292369f72b8
|
/R/methMatrixManipulation.R
|
19c52129b9cafe8a21438d1ad50d0174c9085490
|
[] |
no_license
|
jsemple19/methMatrix
|
0de45df9e5d6c9911d4e1a0356f25e618c55f993
|
9af39085680235b9f380347b52fcc322ade51c0e
|
refs/heads/master
| 2022-08-27T16:32:14.539253
| 2022-08-16T11:56:03
| 2022-08-16T11:56:03
| 165,834,604
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 36,963
|
r
|
methMatrixManipulation.R
|
############ functions for working with methylation matrix lists ##################
# the methylation matrices for individual genes by TSS have the following structure:
# a list (by sample) of lists of matrices (by TSS)
# e.g
# [1] sample1
# [1] TSS1 matrix of reads x Cpositions
# [2] TSS2 matrix of reads x Cpositions
# [2] sample2
# [1] TSS1 matrix of reads x Cpositions
# [2] TSS2 matrix of reads x Cpositions
#
# the matrices contain METHYLATION values: 0 = not methylated, 1 = methylated.
# to avoid confusion, keep them this way and only convert to dSMF (1-methylation) for plotting
# or correlations
#' Convert C position numbering from genomic to relative coordinates
#'
#' @param mat A methylation matrix
#' @param regionGR A genomicRanges object of the region relative to which the new coordinates are caclulated
#' @param invert A logical variable to indicate if the region should be inverted (e.g. if it is on the negative strand). Default: FALSE
#' @return A methylation matrix in which the column names have been changed from absolute genomic positions to relative
#' positions within the genomicRange regionGR
#' @export
getRelativeCoord<-function(mat,regionGR,invert=F){
# converts matrix from absolute genome coordinates to
# relative coordinates within a genomic Range
pos<-as.numeric(colnames(mat))
regionStart<-GenomicRanges::start(regionGR)
regionEnd<-GenomicRanges::end(regionGR)
if (invert==F) {
newPos<-pos-regionStart
colnames(mat)<-as.character(newPos)
} else {
newPos<-regionEnd-pos
colnames(mat)<-newPos
mat<-mat[,order(as.numeric(colnames(mat))),drop=F]
colnames(mat)<-as.character(colnames(mat))
}
return(mat)
}
#' Change the anchor coordinate
#'
#' @param mat A methylation matrix
#' @param anchorCoord The coordinate which will be set as the 0 position for relative coordinates
#' (default=0)
#' @return A methylation matrix in which the column names have been changed to indicate relative position
#' with reference to the anchor coordinate. e.g. a 500bp matrix centered around the TSS can have its column
#' names changed from 0 to 500 range to -250 to 250 range by setting anchorCoord=250.
#' @export
changeAnchorCoord<-function(mat,anchorCoord=0) {
# changes the 0 coordinate position (anchorCoord) of
# a matrix. e.g. sets position 250 to 0 in a 500 region around TSS
# to get +-250 bp around TSS
pos<-as.numeric(colnames(mat))
newPos<-pos-anchorCoord
colnames(mat)<-as.character(newPos)
return(mat)
}
#' Get full matrix of all positions in a window even if no C
#'
#' In order to compare different promoters, we need to create a padded
#' matrix with NAs in positions in between Cs.
#'
#' @param matList A table of paths to matrices which have the same columns
#' @param regionType A name for the type of region the matrices desribe
#' @param winSize The size (in bp) of the window containing the matrices
#' @param workDir The path to working directory
#' @return A table of file paths to padded methylation matrices
#' @export
getFullMatrices<-function(matList,regionType,winSize=500, workDir=".") {
currentSample<-regionGR<-NULL
naRows<-is.na(matList$filename)
matList<-matList[!naRows,]
makeDirs(workDir,paste0("/rds/paddedMats_",regionType))
matrixLog<-matList[,c("filename","sample","region")]
matrixLog$filename<-NA
for (i in 1:nrow(matList)) {
mat<-readRDS(matList$filename[i])
# create a matrix with winSize columns and one row per seq read
Cpos<-colnames(mat)
withinRange<- -winSize/2<=as.numeric(Cpos) & winSize/2>=as.numeric(Cpos)
fullMat<-matrix(data=NaN,nrow=dim(mat)[1],ncol=winSize)
colnames(fullMat)<-c(seq(-winSize/2,-1),seq(1,winSize/2))
fullMat[,Cpos[withinRange]]<-mat[,withinRange]
matName<-paste0(workDir,"/rds/paddedMats_",regionType,"/",currentSample,"_",regionGR$ID,".rds")
saveRDS(fullMat,file=matName)
matrixLog[i,"filename"]<-matName
}
#utils::write.csv(matrixLog,paste0(workDir,"/csv/MatrixLog_paddedMats_",regionType,".csv"), quote=F, row.names=F)
return(matrixLog)
}
#' Convert C position numbering from genomic to relative coordinates for a list of matrices
#'
#' @param matList A table of paths to methylation matrices with names that match the regionGRs object
#' @param regionGRs A genomicRanges object of the regions relative to which the new coordinates are caclulated with a metadata column called "ID" containing names that match the methylation matrices in matList
#' @param regionType A collective name for this list of regions (e.g TSS or amplicons). It will be used in naming the output directories.
#' @param anchorCoord The coordinate which will be set as the 0 position for relative
#' coordinates (default=0)
#' @param workDir path to working directory
#' @return A list of methylation matrices that have been converted from abslute genomic coordinates
#' to relativepositions within the genomicRanges regionGRs. regionGRs on the negative strand will be flipped to be
#' in the forward orientation.
#' @export
getRelativeCoordMats<-function(matList, regionGRs, regionType, anchorCoord=0,workDir=".") {
naRows<-is.na(matList$filename)
matList<-matList[!naRows,]
makeDirs(workDir,paste0("/rds/relCoord_",regionType))
matrixLog<-matList[,c("filename","sample","region")]
matrixLog$filename<-NA
matrixLog$reads<-NA
matrixLog$motifs<-NA
for (i in 1:nrow(matList)) {
print(matList[i,c("sample","region")])
mat<-readRDS(matList$filename[i])
if(sum(dim(mat)==c(0,0))<1) {
regionID<-matList$region[i]
regionGR<-regionGRs[regionGRs$ID==regionID]
newMat<-getRelativeCoord(mat, regionGR,
invert=ifelse(GenomicRanges::strand(regionGR)=="+",F,T))
newMat<-changeAnchorCoord(mat=newMat,anchorCoord=anchorCoord)
} else {
newMat<-mat
}
matName<-paste0(workDir,"/rds/relCoord_",regionType,"/",matList$sample[i],"_",regionGR$ID,".rds")
saveRDS(newMat,file=matName)
matrixLog[i,"filename"]<-matName
matrixLog[i,"reads"]<-dim(newMat)[1]
matrixLog[i,"motifs"]<-dim(newMat)[2]
}
if (! dir.exists(paste0(workDir,"/csv/"))) {
dir.create(paste0(workDir,"/csv/"))
}
utils::write.csv(matrixLog,paste0(workDir,"/csv/MatrixLog_relCoord_",regionType,".csv"), quote=F, row.names=F)
return(matrixLog)
}
#' Calculate methylation frequency at each position for metagene plots
#'
#' @param matList A table of filepaths of methylation matrices with names that match the regionsGRs object
#' @param regionGRs A genomicRanges object of the regions for which aggregate methylation frequency
#' will be calculated. the object must contain a metadata column called "ID" containing names that
#' match the methylation matrices in matList
#' @param minReads The minimal number of reads a matrix must have in order to be used (default=50)
#' @return A long form dataframe with four columns: "position" is the C position within the genomic Ranges,
#' "methFreq" is the frequency of methylation at that position, "ID" is the name of the region, "chr"
#' is the chromosome on which that region is present.
#' @export
getMetaMethFreq<-function(matList,regionGRs,minReads=50) {
naRows<-is.na(matList$filename)
matList<-matList[!naRows,]
first=TRUE
for (i in 1:nrow(matList)) {
mat<-readRDS(matList$filename[i])
if (dim(mat)[1]>minReads) {
vecSummary<-colMeans(mat,na.rm=T)
df<-data.frame("position"=names(vecSummary),"methFreq"=vecSummary,stringsAsFactors=F)
df$ID<-matList$region[i]
df$chr<-as.character(GenomicRanges::seqnames(regionGRs)[match(matList$region[i],regionGRs$ID)])
if (first==TRUE){
methFreqDF<-df
first=FALSE
} else {
methFreqDF<-rbind(methFreqDF,df)
}
}
}
return(methFreqDF)
}
#' Single molecule plot of a methylation matrix
#'
#' @param mat A methylation matrix
#' @param regionName The name of the region the matrix is taken from should match on of the IDs in regionGRs
#' @param regionGRs A genomicRanges object which includes the region which the mat matrix provides the data for.
#' The object must contain a metadata column called "ID" containing names that match the methylation
#' matrices in mat
#' @param featureGRs A genomicRanges object denoting features to be plotted such as the TSS
#' @param myXlab A label for the x axis (default is "CpG/GpC position")
#' @param featureLabel A label for a feature you want to plot, such as the position of the TSS
#' (default="TSS)
#' @param drawArrow Boolean: should the feature be drawn as an arrow or just a line? (default=TRUE)
#' @param title A title for the plot (default will be the name of the region, the chr and strand on which
#' the region is present)
#' @param baseFontSize The base font for the plotting theme (default=12 works well for 4x plots per A4 page)
#' @param maxNAfraction Maximual fraction of CpG/GpC positions that can be undefined (default=0.2)
#' @param segmentSize Length of colour segment denoting methylation site
#' @param colourChoice A list of colours for colour pallette. Must include
#' values for "low", "mid", "high" and "bg" (background) and "lines".
#' @param colourScaleMidpoint Numerical value for middle of colour scale. Useful for Nanopore data where a particular threshold other than 0.5 is used to distinguish methylated from non-methylated sites. (default=0.5).
#' @param doClustering Boolean value to determine if reads should be clustered with heirarchical clustering before plotting (default=T).
#' @return A ggplot2 plot object
#' @export
plotSingleMolecules<-function(mat,regionName, regionGRs, featureGRs=NULL,
myXlab="CpG/GpC position",
featureLabel="TSS", drawArrow=TRUE, title=NULL,
baseFontSize=12, maxNAfraction=0.2, segmentSize=3,
colourChoice=list(low="blue", mid="white",
high="red", bg="white",
lines="black"),
colourScaleMidpoint=0.5, doClustering=T) {
position<-methylation<-molecules<-NULL
### single molecule plot. mat is matrix containing methylation values at different postions
# (columns) in individual reads (rows). regionName is the ID of the amplicon or genomic
# region being plotted. regionGRs is a genomicRanges object containing the region being
# plotted. one of its mcols must have a name "ID" in which the same ID as in regionName
# appears. featureGRs is genomic ranges object for plotting location of some feature in
# the region, such as the TSS. myXlab is the X axis label. featureLabel is the label for
# the type of feature that will be plotted underneath the feature
tooManyNAs<-rowMeans(is.na(mat))>maxNAfraction
mat<-mat[!tooManyNAs,]
if (!is.null(dim(mat)) & any(dim(mat)[1]>10)) {
regionGR<-regionGRs[match(regionName,regionGRs$ID)]
if (length(featureGRs)>0) {
featGR<-featureGRs[match(regionName,featureGRs$ID)]
}
na.matrix<-is.na(mat)
mat[na.matrix]<- -1
# try to perform heirarchical clustering
hc <- try(
stats::hclust(stats::dist(apply(mat,2,as.numeric))),
silent = TRUE)
mat[na.matrix]<-NA
if (class(hc) == "try-error" | !doClustering ) {
df<-as.data.frame(mat,stringsAsFactors=F)
print("hclust not performed. Matrix dim: ")
print(dim(mat))
} else {
df<-as.data.frame(mat[hc$order,], stringsAsFactors=F)
}
reads<-row.names(df)
d<-tidyr::gather(df,key=position,value=methylation)
d$molecules<-seq_along(reads)
#d$methylation<-as.character(d$methylation)
d$position<-as.numeric(d$position)
if (is.null(title)) {
strandInfo<-ifelse(GenomicRanges::strand(featGR)!=
GenomicRanges::strand(regionGR),
paste0("reg: ", GenomicRanges::strand(regionGR),
"ve, ",featureLabel,": ",
GenomicRanges::strand(featGR),"ve strand"),
paste0(GenomicRanges::strand(regionGR),"ve strand"))
title=paste0(regionName, ": ", GenomicRanges::seqnames(regionGR),
" ", strandInfo)
}
scaleFactor<-(GenomicRanges::width(regionGR)/500)
p<-ggplot2::ggplot(d,ggplot2::aes(x=position,y=molecules)) +
ggplot2::geom_tile(ggplot2::aes(width=segmentSize*scaleFactor,
fill=methylation),
alpha=0.8) +
ggplot2::scale_fill_gradient2(low=colourChoice$low, mid=colourChoice$mid,
high=colourChoice$high,
midpoint=colourScaleMidpoint,
na.value=colourChoice$bg,
breaks=c(0,1),
labels=c("protected","accessible"),
limits=c(0,1), name="dSMF\n\n") +
#ggplot2::scale_fill_manual(values=c("0"="black","1"="grey80"),na.translate=F,na.value="white", labels=c("protected","accessible"),name="dSMF") +
ggplot2::theme_light(base_size=baseFontSize) +
ggplot2::theme(panel.grid.major = ggplot2::element_blank(),
panel.grid.minor = ggplot2::element_blank(),
panel.background=ggplot2::element_rect(fill=colourChoice$bg),
plot.title = ggplot2::element_text(face = "bold",hjust = 0.5),
legend.position="bottom",
legend.key.height = ggplot2::unit(0.2, "cm"),
legend.key.width=ggplot2::unit(0.5,"cm")) +
ggplot2::ggtitle(title) +
ggplot2::xlab(myXlab) +
ggplot2::ylab("Single molecules") +
ggplot2::xlim(GenomicRanges::start(regionGR),
GenomicRanges::end(regionGR)+10)
if(!is.null(featureGRs)) {
p<-p+ggplot2::geom_linerange(ggplot2::aes(
x=GenomicRanges::start(featGR), y=NULL, ymin=0,
ymax=length(reads)+max(3,0.04*length(reads))),
col=colourChoice$lines) +
ggplot2::annotate(geom="text", x=GenomicRanges::start(featGR),
y=-max(2,0.03*length(reads)),
label=featureLabel,color=colourChoice$lines)
if (drawArrow==TRUE) {
p<-p+ggplot2::annotate("segment", x = GenomicRanges::start(featGR),
xend = GenomicRanges::start(featGR)+
20*scaleFactor*ifelse(GenomicRanges::strand(featGR)=="-",
-1,1),
y = length(reads)+max(3,0.04*length(reads)),
yend =length(reads)+max(3,0.04*length(reads)),
colour = colourChoice$lines,
arrow=ggplot2::arrow(length = ggplot2::unit(0.3,"cm")),
size=0.7)
}
}
} else {
p<-NULL
}
return(p)
}
#' Single molecule plot of a methylation matrix with average methylation frequency
#'
#' @param mat A methylation matrix
#' @param regionName The name of the region the matrix is taken from should match on of the IDs in regionGRs
#' @param regionGRs A genomicRanges object which includes the region which the mat matrix provides the data for.
#' The object must contain a metadata column called "ID" containing names that match the methylation
#' matrices in mat
#' @param featureGRs A genomicRanges object denoting features to be plotted such as the TSS
#' @param myXlab A label for the x axis (default is "CpG/GpC position")
#' @param featureLabel A label for a feature you want to plot, such as the position of the TSS
#' (default="TSS)
#' @param drawArrow Boolean: should the feature be drawn as an arrow or just a line? (default=TRUE)
#' @param title A title for the plot (default will be the name of the region, the chr and strand on which
#' the region is present)
#' @param baseFontSize The base font for the plotting theme (default=11 works well for 4x plots per A4 page)
#' @param maxNAfraction Maximual fraction of CpG/GpC positions that can be undefined (default=0.2)
#' @param segmentSize Length of colour segment denoting methylation site
#' @param colourChoice A list of colours for colour pallette. Must include
#' values for "low", "mid", "high" and "bg" (background) and "lines".
#' @param colourScaleMidpoint Numerical value for middle of colour scale. Useful for Nanopore data where a particular threshold other than 0.5 is used to distinguish methylated from non-methylated sites. (default=0.5).
#' @param doClustering Boolean value to determine if reads should be clustered with heirarchical clustering before plotting (default=T).
#' @return A ggplot2 plot object
#' @export
plotSingleMoleculesWithAvr<-function(mat, regionName, regionGRs, featureGRs,
myXlab="CpG/GpC position",
featureLabel="TSS", drawArrow=TRUE,
title=NULL, baseFontSize=11,
maxNAfraction=0.2, segmentSize=3,
colourChoice=list(low="blue", mid="white",
high="red", bg="white",
lines="grey80"),
colourScaleMidpoint=0.5, doClustering=T) {
dSMF<-molecules<-position<-methylation<-NULL
# remove reads with more than maxNAfraction positions with NAs
tooManyNAs<-rowMeans(is.na(mat))>maxNAfraction
mat<-mat[!tooManyNAs,]
if(!is.null(dim(mat)) & any(dim(mat)[1]>10)) {
regionGR<-regionGRs[match(regionName,regionGRs$ID)]
if (length(featureGRs)>0) {
featGR<-featureGRs[match(regionName,featureGRs$ID)]
}
na.matrix<-is.na(mat)
mat[na.matrix]<--1
# try to perform heirarchical clustering
hc <- try(
stats::hclust(stats::dist(apply(mat,2,as.numeric))),
silent = TRUE)
mat[na.matrix]<-NA
if (class(hc) == "try-error" | !doClustering ) {
df<-as.data.frame(mat,stringsAsFactors=F)
print("hclust not performed. Matrix dim: ")
print(dim(mat))
} else {
df<-as.data.frame(mat[hc$order,],stringsAsFactors=F)
}
reads<-row.names(df)
d<-tidyr::gather(df,key=position,value=methylation)
d$molecules<-seq_along(reads)
#d$methylation<-as.character(d$methylation)
d$position<-as.numeric(d$position)
if (is.null(title)) {
strandInfo<-ifelse(GenomicRanges::strand(featGR)!=GenomicRanges::strand(regionGR),
paste0("reg: ",GenomicRanges::strand(regionGR),"ve, ",
featureLabel,": ",GenomicRanges::strand(featGR),"ve strand"),
paste0(GenomicRanges::strand(regionGR),"ve strand"))
title=paste0(regionName, ": ",GenomicRanges::seqnames(regionGR)," ",strandInfo)
}
scaleFactor<-(GenomicRanges::width(regionGR)/500)
dAvr<-data.frame(position=as.numeric(colnames(df)),
dSMF=1-colSums(df>=colourScaleMidpoint,
na.rm=T)/length(reads))
# average plot
p1<-ggplot2::ggplot(dAvr,ggplot2::aes(x=position,y=dSMF,group=1)) +
ggplot2::geom_point(size=1/scaleFactor)+
ggplot2::geom_line(size=1,show.legend=F) +
ggplot2::guides(fill=FALSE, color=FALSE) +
ggplot2::theme_light(base_size=baseFontSize) +
ggplot2::ylab("Mean dSMF") +
ggplot2::theme(axis.title.x = ggplot2::element_blank(),
axis.text.x = ggplot2::element_blank(),
plot.background = ggplot2::element_rect(colour="white")) +
ggplot2::ylim(0,1) +
ggplot2::xlim(GenomicRanges::start(regionGR),GenomicRanges::end(regionGR)+20)
if (!is.null(featureGRs)) { # plot feature if present
p1<-p1 + ggplot2::geom_linerange(ggplot2::aes(x=GenomicRanges::start(featGR),
y=NULL, ymin=0, ymax=1),
col=colourChoice$lines,size=0.7)
if (drawArrow==TRUE) {
p1<-p1+ggplot2::annotate("segment", x = GenomicRanges::start(featGR),
xend = GenomicRanges::start(featGR)+
20*scaleFactor*ifelse(GenomicRanges::strand(featGR)=="-",-1,1),
y = 1, yend = 1, colour = colourChoice$lines, size=0.7,
arrow=ggplot2::arrow(length = ggplot2::unit(0.2, "cm")))
}
}
#single molecule plot
p2<-ggplot2::ggplot(d,ggplot2::aes(x=position,y=molecules)) +
ggplot2::geom_tile(ggplot2::aes(width=segmentSize*scaleFactor,fill=methylation),alpha=0.8) +
ggplot2::scale_fill_gradient2(low=colourChoice$low, mid=colourChoice$mid,
high=colourChoice$high, midpoint=colourScaleMidpoint,
na.value=colourChoice$bg,
breaks=c(0,1), labels=c("protected","accessible"),
limits=c(0,1), name="dSMF\n\n") +
ggplot2::theme_light(base_size=baseFontSize) +
ggplot2::theme(panel.grid.major = ggplot2::element_blank(),
panel.grid.minor = ggplot2::element_blank(),
panel.background=ggplot2::element_rect(fill=colourChoice$bg),
plot.title = ggplot2::element_blank(),
legend.position="bottom",
legend.key.height = ggplot2::unit(0.2, "cm"),
legend.key.width=ggplot2::unit(0.5,"cm")) +
ggplot2::xlab(myXlab) +
ggplot2::ylab("Single molecules") +
ggplot2::xlim(GenomicRanges::start(regionGR),GenomicRanges::end(regionGR)+20)
if (length(featureGRs)>0) { # plot feature if present
p2<-p2+ggplot2::geom_linerange(ggplot2::aes(x=GenomicRanges::start(featGR),
y=NULL, ymin=0,
ymax=length(reads)+max(3,0.04*length(reads))),
col=colourChoice$lines) +
ggplot2::annotate(geom="text", x=GenomicRanges::start(featGR),
y=-max(2,0.03*length(reads)),
label=featureLabel, color=colourChoice$lines)
if (drawArrow==TRUE) {
p2<-p2+ggplot2::annotate("segment", x = GenomicRanges::start(featGR),
xend = GenomicRanges::start(featGR)+
20*scaleFactor*ifelse(GenomicRanges::strand(featGR)=="-",-1,1),
y = length(reads)+max(3,0.04*length(reads)),
yend =length(reads)+max(3,0.04*length(reads)),
colour = colourChoice$lines, size=0.7,
arrow=ggplot2::arrow(length = ggplot2::unit(0.3, "cm")))
}
}
figure<-ggpubr::ggarrange(p1, p2, heights = c(0.5, 2), ncol = 1, nrow = 2, align = "v")
figure<-ggpubr::annotate_figure(figure, top = ggpubr::text_grob(title, face = "bold"))
} else {
figure=NULL
}
return(figure)
}
#' Plot a list of list of single molecule matrices
#'
#' This function takes a list (by sample) of a list (by genomic region) of
#' methylation matrices and produces single molecule plots for each amplicon with
#' four samples per page.
#'
#' @param allSampleMats A list (by sample) of lists (by regions) of methylation matrices
#' @param samples A list of samples to plot (same as sample names in allSampleMats)
#' @param regionGRs A genomic regions object with all regions for which matrices should be extracted (same as in allSampleMats). The metadata columns must contain a column called "ID" with a unique ID for each region.
#' @param featureGRs A genomic regions object for features (such as TSS) to be plotted. Feature must be identified with the same ID as the regionGRs
#' @param regionType A collective name for this list of regions (e.g TSS or amplicons). It will be used in naming the output directories
#' @param featureLabel A string with a label for the feature to be added to the plot (default="TSS")
#' @param maxNAfraction Maximual fraction of CpG/GpC positions that can be undefined (default=0.2)
#' @param withAvr Boolean value: should single molecule plots be plotted together with the average profile (default=FALSE)
#' @param includeInFileName String to be included at the end of the plot file name, e.g. experiment name (default="")
#' @param drawArrow Boolean: should the feature be drawn as an arrow or just a line? (default=TRUE)
#' @param workDir Path to working directory
#' @return Plots are written to plots directory
#' @export
plotAllMatrices<-function(allSampleMats, samples, regionGRs, featureGRs, regionType,
featureLabel="TSS", maxNAfraction=0.2,withAvr=FALSE,
includeInFileName="", drawArrow=TRUE, workDir=".") {
# convert any factor variables to character
f <- sapply(allSampleMats, is.factor)
allSampleMats[f] <- lapply(allSampleMats[f], as.character)
# remove any regions with no matrix
naRows<-is.na(allSampleMats$filename)
allSampleMats<-allSampleMats[!naRows,]
# get list of all regions in the object
allAmp2plot<-unique(allSampleMats$region)
# plot single molecule matrices on their own
for (i in allAmp2plot) {
if (withAvr==TRUE) {
makeDirs(workDir,paste0("plots/singleMoleculePlotsAvr_",regionType))
} else {
makeDirs(workDir,paste0("plots/singleMoleculePlots_",regionType))
}
plotList=list()
print(paste0("plotting ", i))
currentRegion<-allSampleMats[allSampleMats$region==i,]
for (j in seq_along(currentRegion$sample)) {
mat<-readRDS(allSampleMats[allSampleMats$region==i &
allSampleMats$sample==currentRegion$sample[j],"filename"])
maxReads=10000
if (!is.null(dim(mat))) {
if (dim(mat)[1]>maxReads) { # if matrix contains more than 10000 reads, do a random subsample
set.seed(1)
chooseRows<-sample(1:dim(mat)[1],maxReads)
mat<-mat[chooseRows,]
}
if (withAvr==TRUE) {
p<-plotSingleMoleculesWithAvr(mat=mat, regionName=i, regionGRs=regionGRs,
featureGRs=featureGRs, myXlab="CpG/GpC position",
featureLabel=featureLabel, drawArrow=drawArrow,
title=currentRegion$sample[j], baseFontSize=11,
maxNAfraction=maxNAfraction)
} else {
p<-plotSingleMolecules(mat=mat, regionName=i, regionGRs=regionGRs,
featureGRs=featureGRs, myXlab="CpG/GpC position",
featureLabel=featureLabel, drawArrow=drawArrow,
title=currentRegion$sample[j], baseFontSize=12,
maxNAfraction=maxNAfraction)
}
if (!is.null(p)) {
plotList[[currentRegion$sample[j]]]<-p
}
}
}
if (length(plotList)>0) {
numPages=ceiling(length(plotList)/4)
for (page in 1:numPages) {
regionGR<-regionGRs[match(i,regionGRs$ID)]
featGR<-featureGRs[match(i,featureGRs$ID)]
chr<-GenomicRanges::seqnames(regionGR)
strandInfo<-ifelse(GenomicRanges::strand(featGR)!=GenomicRanges::strand(regionGR),
paste0("reg: ",GenomicRanges::strand(regionGR),"ve, ",
featureLabel,": ",GenomicRanges::strand(featGR),
"ve strand"),
paste0(GenomicRanges::strand(regionGR),"ve strand"))
title<-paste0(i, ": ",chr," ",strandInfo)
spacer<-ifelse(length(includeInFileName)>0,"_","")
toPlot<-c(1:4)+4*(page-1) #get plots for this page
toPlot<-toPlot[toPlot<=length(plotList)] #make sure only valid plot numbers used
mp<-gridExtra::marrangeGrob(grobs=plotList[toPlot],nrow=2,ncol=2,top=title)
if (withAvr==TRUE) {
ggplot2::ggsave(paste0(workDir,"/plots/singleMoleculePlotsAvr_", regionType,
"/", chr,"_",i,spacer,includeInFileName,"_",page,".png"),
plot=mp, device="png", width=20, height=29, units="cm")
} else {
ggplot2::ggsave(paste0(workDir,"/plots/singleMoleculePlots_",regionType,
"/",chr,"_",i, spacer, includeInFileName,"_",page,".png"),
plot=mp, device="png", width=29, height=20, units="cm")
}
}
}
}
}
#' Convert Genomic Ranges to relative cooridnates
#'
#' Convert a normal genomic ranges to one where the start and end are relative to some
#' anchor point - either the middle or the start of the genomic ranges (e.g. -250 to 250, or
#' 0 to 500). The original start end and strand are stored in the metadata.
#' @param grs A GenomicRanges object to be converted to relative coordinates
#' @param winSize The size of the window you wish to create
#' @param anchorPoint One of "middle" or "start": the position from which numbering starts
#' @return A GenomicRanges object with relative coordinate numbering
#' @export
convertGRtoRelCoord<-function(grs,winSize,anchorPoint="middle") {
grsRelCoord<-grs
GenomicRanges::mcols(grsRelCoord)$gnmStart<-GenomicRanges::start(grs)
GenomicRanges::mcols(grsRelCoord)$gnmEnd<-GenomicRanges::end(grs)
GenomicRanges::mcols(grsRelCoord)$gnmStrand<-GenomicRanges::strand(grs)
grsRelCoord<-GenomicRanges::resize(grsRelCoord,width=winSize,fix="center")
GenomicRanges::strand(grsRelCoord)<-"*"
if (anchorPoint=="middle") {
GenomicRanges::start(grsRelCoord)<- -winSize/2
GenomicRanges::end(grsRelCoord)<- winSize/2
} else if (anchorPoint=="start") {
GenomicRanges::start(grsRelCoord)<- 1
GenomicRanges::end(grsRelCoord)<- winSize
} else {
print("anchorPoint must be one of 'middle' or 'start'")
}
return(grsRelCoord)
}
#' Convert relative cooridnates to absolute genomic position
#'
#' Convert genomic ranges where the start and end are
#' relative to some anchor point - either the middle or the start of
#' the genomic ranges (e.g. -250 to 250, or 0 to 500) to a normal Genomic
#' Ranges with absolute genomic positions.
#' @param grsRelCoord A GenomicRanges object with one or more relative
#' coordinate ranges to be converted to absolute genomic positions.
#' @param regionGR A GenomicRanges object for the whole region to which the
#' grsRelCoord are relative to.
#' @param anchorPoint One of "middle" or "start": the position from which numbering starts
#' @return A GenomicRanges object with absolute genomic position
#' @export
convertRelCoordtoGR<-function(grsRelCoord,regionGR,anchorPoint="middle") {
winSize<-NULL
grs<-grsRelCoord
#grsRelCoord<-GenomicRanges::resize(grsRelCoord,width=winSize,fix="center")
GenomicRanges::strand(grs)<-GenomicRanges::strand(regionGR)
if (anchorPoint=="middle") {
GenomicRanges::start(grs)<- GenomicRanges::start(regionGR) +
winSize/2 + GenomicRanges::start(grsRelCoord)
GenomicRanges::end(grs)<- GenomicRanges::start(regionGR) +
winSize/2 + GenomicRanges::end(grsRelCoord)
} else if (anchorPoint=="start") {
GenomicRanges::start(grs)<- GenomicRanges::start(regionGR) - 1 +
GenomicRanges::start(grsRelCoord)
GenomicRanges::end(grs)<- GenomicRanges::start(regionGR) +
GenomicRanges::end(grsRelCoord)
} else {
print("anchorPoint must be one of 'middle' or 'start'")
}
return(grs)
}
#' get average methylation frequency from all matrices
#'
#' In order to do a metagene plot from matrices, the average methylation frequency from all
#' matrices with more reads than minReads is collected into a data frame
#' @param relCoordMats A table of paths to methylation matrices which have been converted to relative coordinates
#' @param samples A list of samples to plot (same as sample names in relCoordMats)
#' @param regionGRs A genomic regions object with all regions for which matrices should be extracted (same as in relCoordMats). The metadata columns must contain a column called "ID" with a unique ID for each region.
#' @param minReads The minimal number of reads required in a matrix for average frequency to be calculated.
#' @return Data frame with average methylation at relative coordinates extracted from all matrices for all samples
#' @export
getAllSampleMetaMethFreq<-function(relCoordMats,samples,regionGRs,minReads=10) {
naRows<-is.na(relCoordMats$filename)
relCoordMats<-relCoordMats[!naRows,]
first<-TRUE
for (i in seq_along(samples)) {
idx<-relCoordMats$sample==samples[i]
metaMethFreqDF<-getMetaMethFreq(matList=relCoordMats[idx,],
regionGRs=regionGRs, minReads=minReads)
print(samples[i])
metaMethFreqDF$sample<-samples[i]
if(first==TRUE) {
allSampleMetaMethFreqDF<-metaMethFreqDF
first<-FALSE
} else {
allSampleMetaMethFreqDF<-rbind(allSampleMetaMethFreqDF,metaMethFreqDF)
}
}
# convert position from factor to numeric
allSampleMetaMethFreqDF$position<-as.numeric(as.character(allSampleMetaMethFreqDF$position))
return(allSampleMetaMethFreqDF)
}
#' Plot metagene by sample
#'
#' Plots metagene methylation frequency from dataframe extracted from matrices.
#' @param metageneDF ata frame with average methylation at relative coordinates extracted from all matrices for all samples
#' @param maxPoints The maximum number of points to plot per sample. To avoid large files with too much overplotting, the defualt limit is set to 10000. Larger dataframes will be randomly sub-sampled
#' @return A ggplot2 plot object
#' @export
plotDSMFmetageneDF<-function(metageneDF,maxPoints=10000) {
position<-methFreq<-NULL
# subsample if too many points
if (nrow(metageneDF)>maxPoints) {
set.seed(1)
idx<-sample(1:nrow(metageneDF),maxPoints)
} else {
idx<-1:nrow(metageneDF)
}
p1<-ggplot2::ggplot(metageneDF[idx,],ggplot2::aes(x=position,y=1-methFreq)) +
ggplot2::theme_light(base_size=16) + ggplot2::ylim(0,1) +
ggplot2::xlab("Position relative to TSS") + ggplot2::ylab("dSMF (1-%methylation)") +
ggplot2::geom_linerange(ggplot2::aes(x=0, y=NULL, ymin=0, ymax=1),color="steelblue",size=1) +
ggplot2::geom_point(alpha=0.1) +
ggplot2::geom_smooth(colour="red",fill="red") +
ggplot2::facet_wrap(~sample)
p2<-ggplot2::ggplot(metageneDF,ggplot2::aes(x=position,y=1-methFreq,colour=sample)) +
ggplot2::theme_light(base_size=16) + ggplot2::ylim(0,1) +
ggplot2::xlab("Position relative to TSS") + ggplot2::ylab("dSMF (1-%methylation)") +
ggplot2::geom_linerange(ggplot2::aes(x=0, y=NULL, ymin=0, ymax=1),color="steelblue",size=1) +
ggplot2::geom_smooth(se=FALSE)
ml <- gridExtra::marrangeGrob(list(p1,p2), nrow=1, ncol=1)
return(ml)
}
#' Merge tables listing methylation matrices for different samples
#'
#' Uses file naming convention of MatrixLog_regionType_sampleName.csv to
#' merge tables of methylation matrices that were processed separately for
#' each sample. The function will merge all the samples listed in samples and
#' then will also delete the original split files
#' @param path Path to working directory in which csv subdirectory exists.
#' @param regionType A collective name for this list of regions (e.g TSS or amplicons). It is used in naming the files
#' @param samples A list of sample names to merge (used in the name of the file)
#' @param deleteSplitFiles Logical value to determine if individual sample files
#' should be deleted (defualt=F).
#' @return Table of merged samples
#' @export
mergeSampleMats<-function(path, regionType, samples, deleteSplitFiles=F) {
allSampleMats<-NULL
for(s in 1:length(samples)){
if(!file.exists(paste0(path,"/csv/MatrixLog_", regionType, "_",
samples[s], ".csv"))){
cat(paste0("File ",path,"/csv/MatrixLog_", regionType, "_", samples[s],
".csv"," not found"),sep="\n")
next()
}
temp<-utils::read.csv(paste0(path,"/csv/MatrixLog_", regionType, "_",
samples[s], ".csv"),header=T,
stringsAsFactors=F)
if(is.null(allSampleMats)) {
allSampleMats<-temp
} else {
allSampleMats<-rbind(allSampleMats,temp)
}
}
utils::write.csv(allSampleMats,paste0(path, "/csv/MatrixLog_", regionType,".csv"),
quote=F, row.names=F)
# delete split files
if(deleteSplitFiles){
for(s in 1:length(samples)){
file.remove(paste0(path, "/csv/MatrixLog_", regionType, "_",samples[s],
".csv"))
file.remove(paste0(path, "/csv/MatrixLog_", regionType, "_", samples[s],
"_log.csv"))
}
}
return(allSampleMats)
}
|
2231a4293ef5cfcaa22324f4bdd0b958043a5ce8
|
fc214123e9c4caca64b4a063e67df94584277771
|
/0_GraphEffects.R
|
180d3e44b60b4bc742e1229d91a1e993f1de5e8a
|
[] |
no_license
|
alexanderm10/canopy_class_climate
|
e6ecc886ba668be3ef969af2227365c3f76d3971
|
ebc780ec0cda97492beb9c531b933ca95cadce42
|
refs/heads/master
| 2021-05-09T20:09:12.334030
| 2020-11-04T20:05:55
| 2020-11-04T20:05:55
| 118,674,596
| 0
| 1
| null | null | null | null |
UTF-8
|
R
| false
| false
| 16,979
|
r
|
0_GraphEffects.R
|
# Making functions so that all models can get graphed using the same parameters
cbbPalette <- c("#009E73", "#e79f00", "#9ad0f3", "#0072B2", "#D55E00")
# plotting the size effect; standard dimensions = 8 tall, 5 wide
plot.size <- function(dat.plot){
ggplot(data=dat.plot[dat.plot$Effect=="dbh.recon",]) +
# ggtitle("Null Model") +
facet_grid(Species~.) +
geom_ribbon(aes(x=x, ymin=lwr.bai*100, ymax=upr.bai*100), alpha=0.5) +
geom_line(aes(x=x, y=mean.bai*100)) +
geom_hline(yintercept=100, linetype="dashed") +
scale_x_continuous(expand=c(0,0)) +
# coord_cartesian(ylim=c(0, 1750)) +
labs(x = expression(bold(paste("DBH (cm)"))), y = expression(bold(paste("Relativized BAI (%)"))))+
theme(axis.line=element_line(color="black"),
panel.grid.major=element_blank(),
panel.grid.minor=element_blank(),
panel.border=element_blank(),
panel.background=element_rect(fill=NA, color="black"),
axis.text.x=element_text(angle=0, color="black", size=10),
axis.text.y=element_text(angle=0, color="black", size=10),
strip.text=element_text(face="bold", size=12),
axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5),
legend.position="top",
legend.key.size = unit(0.75, "cm"),
legend.text = element_text(size=10),
legend.key = element_rect(fill = "white")) +
#guides(color=guide_legend(nrow=1),)+
theme(axis.title.x = element_text(size=12, face="bold"),
axis.title.y= element_text(size=12, face="bold"))+
theme(panel.spacing.x = unit(0.5,"lines"),
panel.spacing.y = unit(0.5,"lines"))
}
plot.year <- function(dat.plot){
ggplot(data=dat.plot[dat.plot$Effect=="Year",]) +
# ggtitle("Null Model") +
facet_wrap(~PlotID, scales="free_y") +
geom_ribbon(aes(x=x, ymin=lwr.bai*100, ymax=upr.bai*100, fill=Species), alpha=0.5) +
geom_line(aes(x=x, y=mean.bai*100, color=Species)) +
geom_hline(yintercept=100, linetype="dashed") +
scale_x_continuous(expand=c(0,0)) +
coord_cartesian(ylim=c(0,300)) +
# coord_cartesian(ylim=c(0, 1750)) +
labs(x = expression(bold(paste("Year"))), y = expression(bold(paste("Relativized BAI (%)")))) +
theme(axis.line=element_line(color="black"),
panel.grid.major=element_blank(),
panel.grid.minor=element_blank(),
panel.border=element_blank(),
panel.background=element_rect(fill=NA, color="black"),
axis.text.x=element_text(angle=-45, hjust=0, color="black", size=10),
axis.text.y=element_text(angle=0, color="black", size=10),
strip.text=element_text(face="bold", size=10),
axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5),
legend.position="top",
legend.key.size = unit(0.75, "cm"),
legend.text = element_text(size=10),
legend.key = element_rect(fill = "white")) +
#guides(color=guide_legend(nrow=1),)+
theme(axis.title.x = element_text(size=12, face="bold"),
axis.title.y= element_text(size=12, face="bold"))+
theme(panel.spacing.x = unit(0.5,"lines"),
panel.spacing.y = unit(0.5,"lines"),
plot.margin=unit(c(0.5, 2, 0.5, 0.5), "lines"))
}
# ---------------------------------------------
# The big one: 3-panel climate effects
# ---------------------------------------------
plot.climate <- function(dat.plot, canopy=F, species=F, ...){
plot.base <- ggplot(data=dat.plot[dat.plot$Effect%in%c("tmean", "precip", "vpd.max"),]) +
# facet_grid(.~Effect) +
coord_cartesian(ylim=c(50, 200)) +
geom_hline(yintercept=100, linetype="dashed") +
scale_y_continuous(limits=c(min(dat.plot$lwr.bai[dat.plot$Effect %in% c("tmean", "precip", "vpd.max")]*100, na.rm=T), max(dat.plot$upr.bai[dat.plot$Effect %in% c("tmean", "precip", "vpd.max")]*100, na.rm=T))) +
theme(axis.line=element_line(color="black"),
panel.grid.major=element_blank(),
panel.grid.minor=element_blank(),
panel.border=element_blank(),
panel.background=element_rect(fill=NA, color="black"),
axis.ticks.length = unit(-0.5, "lines"),
axis.text.x = element_text(margin=unit(c(1,1,1,1), "lines"), color="black", size=10),
axis.text.y = element_text(margin=unit(c(1,1,1,1), "lines"), color="black", size=10),
strip.text=element_text(face="bold", size=18),
axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5),
legend.position="top",
# legend.key.size = unit(0.75, "cm"),
legend.text = element_text(size=12),
legend.key = element_rect(fill = "white")) +
#guides(color=guide_legend(nrow=1),)+
theme(axis.title.x = element_text(size=12, face="bold"),
axis.title.y= element_text(size=12, face="bold")) +
theme(panel.spacing.x = unit(0.5,"lines"),
panel.spacing.y = unit(0.5,"lines"),
strip.text.x = element_blank(),
plot.background = element_rect(fill=NA, color=NA))
if(species){
plot.base <- plot.base + facet_grid(Species ~ Effect)
} else {
plot.base <- plot.base + facet_grid(.~Effect)
}
if(canopy){
plot.base <- plot.base +
geom_ribbon(aes(x=x, ymin=lwr.bai*100, ymax=upr.bai*100, fill=Canopy.Class), alpha=0.5) +
geom_line(aes(x=x, y=mean.bai*100, color=Canopy.Class)) +
scale_fill_manual(values=c("#E69F00","#009E73", "#0072B2"))+
scale_color_manual(values=c("#E69F00","#009E73", "#0072B2")) +
theme(legend.title = element_blank())
} else {
plot.base <- plot.base +
geom_ribbon(aes(x=x, ymin=lwr.bai*100, ymax=upr.bai*100), alpha=0.5) +
geom_line(aes(x=x, y=mean.bai*100))
}
plot.tmean <- plot.base %+% subset(dat.plot, Effect=="tmean") +
labs(x = expression(bold(paste("Temperature ("^"o", "C)"))), y = expression(bold(paste("Relativized BAI (%)")))) +
guides(fill=F, color=F) +
theme(strip.text.y = element_blank(),
axis.text.x = element_text(margin=unit(c(1,1,1,1), "lines"), color="black", size=10),
axis.title.x = element_text(margin=unit(c(0,0,0,0), "lines"), color="black", size=12),
plot.margin = unit(c(3.75,0.5, 0.5, 1), "lines"))
if(canopy){
plot.precip <- plot.base %+% subset(dat.plot, Effect=="precip") +
labs(x = expression(bold(paste("Precipitation (mm)"))), y = element_blank()) +
theme(axis.text.y=element_blank(),
axis.ticks.y=element_line(unit(-0.5, units="lines")),
strip.text.y = element_blank(),
plot.margin = unit(c(1,0.5, 0.5, 0.5), "lines"))
} else {
plot.precip <- plot.base %+% subset(dat.plot, Effect=="precip") +
labs(x = expression(bold(paste("Precipitation (mm)"))), y = element_blank()) +
theme(axis.text.y=element_blank(),
axis.ticks.y=element_line(unit(-0.5, units="lines")),
strip.text.y = element_blank(),
plot.margin = unit(c(3.75,0.5, 0.5, 0.5), "lines"))
}
plot.vpd <- plot.base %+% subset(dat.plot, Effect=="vpd.max") +
labs(x = expression(bold(paste("VPD (kPa)"))), y = element_blank()) +
guides(fill=F, color=F) +
theme(axis.text.y=element_blank(),
axis.ticks.y=element_line(unit(-0.5, units="lines")),
plot.margin = unit(c(3.75,1, 0.5, 0.5), "lines"))
cowplot::plot_grid(plot.tmean, plot.precip, plot.vpd, nrow=1, rel_widths = c(1.5, 1, 1.25))
}
plot.climate.site <- function(dat.plot, canopy=T, species=F, ...){
if(species) stop("We're using the function that plots sites. Can't do both species & site!")
plot.base <- ggplot(data=dat.plot[dat.plot$Effect%in%c("tmean", "precip", "vpd.max"),]) +
facet_grid(Site.Code~Effect) +
geom_hline(yintercept=100, linetype="dashed") +
scale_y_continuous(limits=c(min(dat.plot$lwr.bai[dat.plot$Effect %in% c("tmean", "precip", "vpd.max")]*100, na.rm=T), max(dat.plot$upr.bai[dat.plot$Effect %in% c("tmean", "precip", "vpd.max")]*100, na.rm=T))) +
coord_cartesian(ylim=c(50, 200)) +
theme(axis.line=element_line(color="black"),
panel.grid.major=element_blank(),
panel.grid.minor=element_blank(),
panel.border=element_blank(),
panel.background=element_rect(fill=NA, color="black"),
axis.ticks.length = unit(-0.5, "lines"),
axis.text.x = element_text(margin=unit(c(1,1,1,1), "lines"), color="black", size=10),
axis.text.y = element_text(margin=unit(c(1,1,1,1), "lines"), color="black", size=10),
strip.text=element_text(face="bold", size=18),
axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5),
legend.position="top",
# legend.key.size = unit(0.75, "cm"),
legend.text = element_text(size=12),
legend.key = element_rect(fill = "white")) +
#guides(color=guide_legend(nrow=1),)+
theme(axis.title.x = element_text(size=12, face="bold"),
axis.title.y= element_text(size=12, face="bold")) +
theme(panel.spacing.x = unit(0.5,"lines"),
panel.spacing.y = unit(0.5,"lines"),
strip.text.x = element_blank(),
plot.background = element_rect(fill=NA, color=NA))
if(canopy){
plot.base <- plot.base +
geom_ribbon(aes(x=x, ymin=lwr.bai*100, ymax=upr.bai*100, fill=Canopy.Class), alpha=0.5) +
geom_line(aes(x=x, y=mean.bai*100, color=Canopy.Class)) +
scale_fill_manual(values=c("#E69F00","#009E73", "#0072B2"))+
scale_color_manual(values=c("#E69F00","#009E73", "#0072B2")) +
theme(legend.title = element_blank())
} else {
plot.base <- plot.base +
geom_ribbon(aes(x=x, ymin=lwr.bai*100, ymax=upr.bai*100), alpha=0.5) +
geom_line(aes(x=x, y=mean.bai*100))
}
plot.tmean <- plot.base %+% subset(dat.plot, Effect=="tmean") +
labs(x = expression(bold(paste("Temperature ("^"o", "C)"))), y = expression(bold(paste("Relativized BAI (%) (%)")))) +
guides(fill=F, color=F) +
theme(strip.text.y = element_blank(),
axis.text.x = element_text(margin=unit(c(1,1,1,1), "lines"), color="black", size=10),
axis.title.x = element_text(margin=unit(c(0,0,0,0), "lines"), color="black", size=12),
plot.margin = unit(c(3.75,0.5, 0.5, 1), "lines"))
if(canopy){
plot.precip <- plot.base %+% subset(dat.plot, Effect=="precip") +
labs(x = expression(bold(paste("Precipitation (mm)"))), y = element_blank()) +
theme(axis.text.y=element_blank(),
axis.ticks.y=element_line(unit(-0.5, units="lines")),
strip.text.y = element_blank(),
plot.margin = unit(c(1,0.5, 0.5, 0.5), "lines"))
} else {
plot.precip <- plot.base %+% subset(dat.plot, Effect=="precip") +
labs(x = expression(bold(paste("Precipitation (mm)"))), y = element_blank()) +
theme(axis.text.y=element_blank(),
axis.ticks.y=element_line(unit(-0.5, units="lines")),
strip.text.y = element_blank(),
plot.margin = unit(c(3.75,0.5, 0.5, 0.5), "lines"))
}
plot.vpd <- plot.base %+% subset(dat.plot, Effect=="vpd.max") +
labs(x = expression(bold(paste("VPD (kPa)"))), y = element_blank()) +
guides(fill=F, color=F) +
theme(axis.text.y=element_blank(),
axis.ticks.y=element_line(unit(-0.5, units="lines")),
plot.margin = unit(c(3.75,1, 0.5, 0.5), "lines"))
cowplot::plot_grid(plot.tmean, plot.precip, plot.vpd, nrow=1, rel_widths = c(1.5, 1, 1.25))
}
# ---------------------------------------------
# ---------------------------------------------
# The big one: 3-panel climate effects for leave-one-out analysis
# ---------------------------------------------
plot.climate.sites <- function(dat.plot, canopy=F, species=NULL, panel="sites", ...){
if(!canopy) panel="sites"
if(is.null(species)) species <- unique(dat.plot$Species)
plot.base <- ggplot(data=dat.plot[dat.plot$Species %in% species & dat.plot$Effect%in%c("tmean", "precip", "vpd.max"),]) +
# facet_grid(.~Effect) +
coord_cartesian(ylim=c(50, 200)) +
geom_hline(yintercept=100, linetype="dashed") +
scale_y_continuous(limits=c(min(dat.plot$lwr.bai[dat.plot$Effect %in% c("tmean", "precip", "vpd.max")]*100, na.rm=T), max(dat.plot$upr.bai[dat.plot$Effect %in% c("tmean", "precip", "vpd.max")]*100, na.rm=T))) +
theme(axis.line=element_line(color="black"),
panel.grid.major=element_blank(),
panel.grid.minor=element_blank(),
panel.border=element_blank(),
panel.background=element_rect(fill=NA, color="black"),
axis.ticks.length = unit(-0.5, "lines"),
axis.text.x = element_text(margin=unit(c(1,1,1,1), "lines"), color="black", size=10),
axis.text.y = element_text(margin=unit(c(1,1,1,1), "lines"), color="black", size=10),
strip.text=element_text(face="bold", size=18),
axis.line.x = element_line(color="black", size = 0.5),
axis.line.y = element_line(color="black", size = 0.5),
legend.position="top",
# legend.key.size = unit(0.75, "cm"),
legend.text = element_text(size=12),
legend.key = element_rect(fill = "white")) +
#guides(color=guide_legend(nrow=1),)+
theme(axis.title.x = element_text(size=12, face="bold"),
axis.title.y= element_text(size=12, face="bold")) +
theme(panel.spacing.x = unit(0.5,"lines"),
panel.spacing.y = unit(0.5,"lines"),
strip.text.x = element_blank(),
plot.background = element_rect(fill=NA, color=NA))
if(panel=="sites"){
if(!canopy){
plot.base <- plot.base + facet_grid(Species ~ Effect)
} else {
plot.base <- plot.base + facet_grid(Canopy.Class ~ Effect)
}
plot.base <- plot.base +
geom_ribbon(aes(x=x, ymin=lwr.bai*100, ymax=upr.bai*100, fill=Site), alpha=0.5) +
geom_line(aes(x=x, y=mean.bai*100, color=Site)) +
scale_fill_manual(name="SiteOut", values=c("HO"="#792427FF", "GB"="#633D43FF", "RH"="#4E565FFF", "GE"="#36727EFF", "PS"="#438990FF", "NR"="#739B96FF", "HF"="#A2AC9CFF", "LF"="#D1BDA2FF"))+
scale_color_manual(name="SiteOut", values=c("HO"="#792427FF", "GB"="#633D43FF", "RH"="#4E565FFF", "GE"="#36727EFF", "PS"="#438990FF", "NR"="#739B96FF", "HF"="#A2AC9CFF", "LF"="#D1BDA2FF"))+
theme(legend.title = element_blank())
} else {
# Panel shows comparisons of canopy classes
plot.base <- plot.base + facet_grid(Site ~ Effect) +
geom_ribbon(aes(x=x, ymin=lwr.bai*100, ymax=upr.bai*100, fill=Canopy.Class), alpha=0.5) +
geom_line(aes(x=x, y=mean.bai*100, color=Canopy.Class)) +
scale_fill_manual(values=c(Overstory="#E69F00","Middle"="#009E73", "Understory"="#0072B2"))+
scale_color_manual(values=c(Overstory="#E69F00","Middle"="#009E73", "Understory"="#0072B2")) +
theme(legend.title = element_blank())
} # End sites-based panels or not
# test <- subset(dat.plot, Effect=="tmean", Species==species)
plot.tmean <- plot.base %+% dat.plot[dat.plot$Effect=="tmean" & dat.plot$Species==species, ] +
labs(x = expression(bold(paste("Temperature ("^"o", "C)"))), y = expression(bold(paste("Relativized BAI (%)")))) +
guides(fill=F, color=F) +
theme(strip.text.y = element_blank(),
axis.text.x = element_text(margin=unit(c(1,1,1,1), "lines"), color="black", size=10),
axis.title.x = element_text(margin=unit(c(0,0,0,0), "lines"), color="black", size=12),
plot.margin = unit(c(4.75,0.5, 0.5, 1), "lines"))
plot.precip <- plot.base %+% dat.plot[dat.plot$Effect=="precip" & dat.plot$Species==species, ] +
labs(x = expression(bold(paste("Precipitation (mm)"))), y = element_blank()) +
theme(axis.text.y=element_blank(),
axis.ticks.y=element_line(unit(-0.5, units="lines")),
strip.text.y = element_blank())
if(!panel=="sites"){
plot.precip <- plot.precip + theme(plot.margin = unit(c(2,0.5, 0.5, 0.5), "lines"))
} else {
plot.precip <- plot.precip + theme(plot.margin = unit(c(0.8,0.5, 0.5, 0.5), "lines"))
} # End setting margins based on color level
plot.vpd <- plot.base %+% dat.plot[dat.plot$Effect=="vpd.max" & dat.plot$Species==species, ] +
labs(x = expression(bold(paste("VPD (kPa)"))), y = element_blank()) +
guides(fill=F, color=F) +
theme(axis.text.y=element_blank(),
axis.ticks.y=element_line(unit(-0.5, units="lines")),
plot.margin = unit(c(4.75,1, 0.5, 0.5), "lines"))
cowplot::plot_grid(plot.tmean, plot.precip, plot.vpd, nrow=1, rel_widths = c(1.5, 1, 1.25))
} # End function
# ---------------------------------------------
|
5f075ed8daf8b88d2a27da3d3b5c18c2531882b4
|
06c93d110bc8441b6fafba6e2df030224e87cb2b
|
/man/create_course_assigment.Rd
|
333d36c826eff3b853c9b0beaaccd1e1178a09a6
|
[] |
no_license
|
wsphd/rcanvas
|
57162905a0d860b80ac0cc548cab259a72a8fd6e
|
ad8929935f410d196a9891a07f537e4b6c8d3ab2
|
refs/heads/master
| 2020-03-25T05:49:58.661865
| 2018-08-03T20:04:54
| 2018-08-03T20:04:54
| 143,467,846
| 0
| 0
| null | 2018-08-03T19:51:25
| 2018-08-03T19:51:25
| null |
UTF-8
|
R
| false
| true
| 6,207
|
rd
|
create_course_assigment.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/uploads.R
\name{create_course_assigment}
\alias{create_course_assigment}
\title{Create a course assignment}
\usage{
create_course_assigment(course_id, name, position = NULL,
submission_types = NULL, allowed_extensions = NULL,
turnitin_enabled = NULL, vericite_enabled = NULL,
turnitin_settings = NULL, integration_data = NULL,
integration_id = NULL, peer_reviews = NULL,
automatic_peer_reviews = NULL, notify_of_update = NULL,
group_category_id = NULL, grade_group_students_individually = NULL,
external_tool_tag_attributes = NULL, points_possible = NULL,
grading_type = NULL, due_at = NULL, lock_at = NULL, unlock_at = NULL,
description = NULL, assignment_group_id = NULL, muted = NULL,
assignment_overrides = NULL, only_visible_to_overrides = NULL,
published = NULL, grading_standard_id = NULL,
omit_from_final_grade = NULL, quiz_lti = NULL)
}
\arguments{
\item{course_id}{a valid course id}
\item{name}{the assignment name (only parameter required)}
\item{position}{integer - The position of this assignment in the group when displaying assignment lists.}
\item{submission_types}{string - List of supported submission types for the assignment. Unless the assignment is allowing online submissions, the array should only have one element. Options: online_quiz, none, on_paper, discussion_topic, external_tool, online_upload, online_text_entry, online_url, media_recording}
\item{allowed_extensions}{Allowed extensions if submission_types includes “online_upload”. E.g. "docx", "png".}
\item{turnitin_enabled}{boolean - Only applies when the Turnitin plugin is enabled for a course and the submission_types array includes “online_upload”. Toggles Turnitin submissions for the assignment. Will be ignored if Turnitin is not available for the course.}
\item{vericite_enabled}{boolean - Only applies when the VeriCite plugin is enabled for a course and the submission_types array includes “online_upload”. Toggles VeriCite submissions for the assignment. Will be ignored if VeriCite is not available for the course.}
\item{turnitin_settings}{string - Settings to send along to turnitin. See Assignment object definition for format.}
\item{integration_data}{string - Data used for SIS integrations. Requires admin-level token with the “Manage SIS” permission. JSON string required.}
\item{integration_id}{string - Unique ID from third party integrations}
\item{peer_reviews}{boolean - If submission_types does not include external_tool,discussion_topic, online_quiz, or on_paper, determines whether or not peer reviews will be turned on for the assignment.}
\item{automatic_peer_reviews}{boolean - Whether peer reviews will be assigned automatically by Canvas or if teachers must manually assign peer reviews. Does not apply if peer reviews are not enabled.}
\item{notify_of_update}{boolean - If true, Canvas will send a notification to students in the class notifying them that the content has changed.}
\item{group_category_id}{integer - If present, the assignment will become a group assignment assigned to the group.}
\item{grade_group_students_individually}{boolean - If this is a group assignment, teachers have the options to grade students individually. If false, Canvas will apply the assignment's score to each member of the group. If true, the teacher can manually assign scores to each member of the group.}
\item{external_tool_tag_attributes}{string - Hash of external tool parameters if submission_types is external_tool. See Assignment object definition for format.}
\item{points_possible}{number - The maximum points possible on the assignment.}
\item{grading_type}{string - The strategy used for grading the assignment. The assignment defaults to “points” if this field is omitted. Options: pass_fail, percent, letter_grade, gpa_scale, points}
\item{due_at}{datetime - The day/time the assignment is due. Must be between the lock dates if there are lock dates. Accepts times in ISO 8601 format, e.g. 2014-10-21T18:48:00Z.}
\item{lock_at}{datetime - The day/time the assignment is locked after. Must be after the due date if there is a due date. Accepts times in ISO 8601 format, e.g. 2014-10-21T18:48:00Z.}
\item{unlock_at}{datetime - The day/time the assignment is unlocked. Must be before the due date if there is a due date. Accepts times in ISO 8601 format, e.g. 2014-10-21T18:48:00Z.}
\item{description}{string - The assignment's description, supports HTML.}
\item{assignment_group_id}{number - The assignment group id to put the assignment in. Defaults to the top assignment group in the course.}
\item{muted}{boolean - Whether this assignment is muted. A muted assignment does not send change notifications and hides grades from students. Defaults to false.}
\item{assignment_overrides}{List of overrides for the assignment.}
\item{only_visible_to_overrides}{boolean - Whether this assignment is only visible to overrides (Only useful if 'differentiated assignments' account setting is on)}
\item{published}{boolean - Whether this assignment is published. (Only useful if 'draft state' account setting is on) Unpublished assignments are not visible to students.}
\item{grading_standard_id}{integer - The grading standard id to set for the course. If no value is provided for this argument the current grading_standard will be un-set from this course. This will update the grading_type for the course to letter_grade' unless it is already 'gpa_scale'.}
\item{omit_from_final_grade}{boolean - Whether this assignment is counted towards a student's final grade.}
\item{quiz_lti}{boolean - Whether this assignment should use the Quizzes 2 LTI tool. Sets the submission type to 'external_tool' and configures the external tool attributes to use the Quizzes 2 LTI tool configured for this course. Has no effect if no Quizzes 2 LTI tool is configured.}
}
\description{
Create a course assignment
}
\examples{
create_course_assignment(course_id = 432432, name = "Challenging Assignment")
create_course_assignment(course_id = 3432432, name = "R Packages, Review", peer_reviews = TRUE, points_possible = 100, omit_from_final_grade = TRUE)
}
|
467a5cb825e22f212b5a78ea30c33d81750a7033
|
1f6d79658ce351eafa3bf83cf38949d82b58de2f
|
/man/diffnet_check_attr_class.Rd
|
036294892729f14b1260efbf1df6e3d7e2f102ad
|
[
"MIT"
] |
permissive
|
USCCANA/netdiffuseR
|
3dd061f8b9951f7bdc5ec69cded73144f6a63cf7
|
7c5c9a7d4a8120491bfd44d6e307bdb5b66c18ae
|
refs/heads/master
| 2023-09-01T08:26:19.951911
| 2023-08-30T15:44:09
| 2023-08-30T15:44:09
| 28,208,077
| 85
| 23
|
NOASSERTION
| 2020-03-14T00:54:59
| 2014-12-19T00:44:59
|
R
|
UTF-8
|
R
| false
| true
| 728
|
rd
|
diffnet_check_attr_class.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/diffnet-indexing.r
\name{diffnet_check_attr_class}
\alias{diffnet_check_attr_class}
\title{Infer whether \code{value} is dynamic or static.}
\usage{
diffnet_check_attr_class(value, meta)
}
\arguments{
\item{value}{Either a matrix, data frame or a list. Attribute values.}
\item{meta}{A list. A diffnet object's meta data.}
}
\value{
The value object either as a data frame (if static) or as a list
of data frames (if dynamic). If \code{value} does not follows the permitted
types of \code{\link{diffnet_index}}, then returns with error.
}
\description{
Intended for internal use only, this function is used in \code{\link{diffnet_index}}
methods.
}
|
42e69cb0f2aed966673789ba384d84b1758e4571
|
e8577e571531992fa56a9173a19c55529716502b
|
/tests/testthat/test-golem.R
|
7a39a39b69320843a697c95583a75bd0ae150bfe
|
[
"MIT"
] |
permissive
|
DivadNojnarg/packer
|
bcd2c3a13da1308761fcef6cc124aeabfc478b2b
|
332e01964835bf4cdf3d5dfc1eb0f2c43d3db107
|
refs/heads/master
| 2023-08-29T14:59:35.402851
| 2021-10-17T12:27:30
| 2021-10-17T12:27:30
| 416,646,602
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 2,166
|
r
|
test-golem.R
|
source("../fns.R")
skip_on_cran()
test_that("Golem Bare", {
# keep working directory
wd <- getwd()
# test bare
pkg <- create_tmp_golem()
setwd(pkg)
on.exit({
setwd(wd)
delete_tmp_package(pkg)
})
expect_output(scaffold_golem(edit = FALSE))
expect_error(scaffold_golem(edit = FALSE))
expect_message(bundle_dev())
expect_message(add_plugin_html(use_pug = TRUE))
expect_message(add_plugin_prettier())
expect_message(add_plugin_eslint())
expect_message(add_plugin_jsdoc(FALSE))
add_jsdoc_tutorial("xxx", FALSE)
expect_message(add_plugin_workbox())
})
test_that("Golem CDN", {
# keep working directory
wd <- getwd()
# test react
pkg <- create_tmp_golem()
setwd(pkg)
expect_output(scaffold_golem(react = TRUE, edit = FALSE))
expect_message(bundle())
expect_message(use_loader_mocha(FALSE))
setwd(wd)
delete_tmp_package(pkg)
# test vue
pkg <- create_tmp_golem()
setwd(pkg)
expect_output(scaffold_golem(vue = TRUE, edit = FALSE))
expect_message(bundle())
expect_message(add_plugin_clean())
setwd(wd)
delete_tmp_package(pkg)
})
test_that("Golem no CDN", {
# keep working directory
wd <- getwd()
# test react
pkg <- create_tmp_golem()
setwd(pkg)
expect_output(scaffold_golem(react = TRUE, use_cdn = FALSE, edit = FALSE))
expect_message(bundle())
expect_message(npm_console())
expect_message(npm_run("production"))
setwd(wd)
delete_tmp_package(pkg)
# test vue
pkg <- create_tmp_golem()
setwd(pkg)
expect_output(scaffold_golem(vue = TRUE, use_cdn = FALSE, edit = FALSE))
expect_message(bundle())
setwd(wd)
delete_tmp_package(pkg)
# test framework7
pkg <- create_tmp_golem()
setwd(pkg)
expect_output(scaffold_golem(framework7 = TRUE, edit = FALSE))
expect_message(bundle())
setwd(wd)
delete_tmp_package(pkg)
})
test_that("Golem F7", {
# keep working directory
wd <- getwd()
# test bare
pkg <- create_tmp_golem()
setwd(pkg)
on.exit({
setwd(wd)
delete_tmp_package(pkg)
})
expect_output(scaffold_golem(framework7 = TRUE, edit = FALSE))
expect_error(scaffold_golem(edit = FALSE))
expect_message(bundle_dev())
})
|
affa1b8754127a55c4d0e31899f70c5a9c75cae2
|
1ea6b75a27e313ec0a0386e28352f390a6915677
|
/Examples/4_DefaultCreditCard.r
|
1740766c57eed3f983e319e08ecaa2da37014a40
|
[] |
no_license
|
NQuinn27/NCIY4_R_Labs
|
7cb160943d64868282e9df9e7ad6fd75f3d560bf
|
ce9f062b5bc9ed022fc59344ed9beaa4ddd9c2ff
|
refs/heads/master
| 2021-01-11T02:19:03.436797
| 2016-10-15T09:46:10
| 2016-10-15T09:46:10
| 70,979,428
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 2,660
|
r
|
4_DefaultCreditCard.r
|
install.packages(c("e1071", "C50", "ggplot2", "hexbin","descr", "caret", "e1071"))
library(e1071)
library(hexbin)
library(ggplot2)
library(caret)
library(descr)
library(C50)
setwd("Developer/NCI/Data Application Design")
# First remove first row (i.e., X1, X2 etc.) and export the data as .csv from Excel
data <- read.csv("default of credit card clients.csv", stringsAsFactors=F)
data <- data[-1]
str(data)
View(data)
summary(data)
#tranform the #SEX column from numeric to a Factor (vector with associated labels)
data$SEX <- factor(data$SEX, levels=c(1, 2), labels=c("M", "F"))
summary(data)
table(data$EDUCATION, useNA='ifany')
table(data$MARRIAGE, useNA='ifany')
data$EDUCATION <- factor(data$EDUCATION, levels=c(0, 1, 2, 3, 4, 5, 6), labels=c(NA, "GS", "UNI", "HS", "O1", "O2", "O3"))
data$MARRIAGE <- factor(data$MARRIAGE, levels=c(0, 1, 2, 3), labels=c(NA, "M", "S", "O"))
data$default.payment.next.month <- factor(data$default.payment.next.month, levels=c(0, 1), labels=c("N", "Y"))
str(data)
summary(data)
# Descriptive statistics and plots
mean(data$LIMIT_BAL)
summary(data$LIMIT_BAL)
median(data$LIMIT_BAL)
sd(data$LIMIT_BAL)
IQR(data$LIMIT_BAL)
mad(data$LIMIT_BAL)
boxplot(data$LIMIT_BAL)
hist(data$LIMIT_BAL, freq=F)
lines(density(data$LIMIT_BAL), lwd=3, col="blue")
ggplot(data, aes(x=LIMIT_BAL, y=AGE)) + stat_binhex(colour="white") +
theme_bw() + scale_fill_gradient(low="white", high="blue") +
labs(x="Limit Balance", y="Age")
ggplot(data, aes(x=LIMIT_BAL, y=AGE)) + stat_binhex(colour="white") +
theme_bw() + scale_fill_gradient(low="white", high="blue") +
labs(x="Limit Balance", y="Age") +
facet_wrap("EDUCATION")
CrossTable(data$EDUCATION, data$default.payment.next.month, prop.c=F, prop.t=F, prop.chisq=F)
boxplot(LIMIT_BAL ~ EDUCATION, data=data)
ggplot(data=data, aes(EDUCATION, LIMIT_BAL)) +
geom_violin(fill="lightblue") +
geom_boxplot( alpha=.2)
# Randomise data
data_rand <- data[order(runif(10000)), ]
summary(data$LIMIT_BAL)
summary(data_rand$LIMIT_BAL)
# Create test and train subsets
train <- data_rand[1:9000, ]
test <- data_rand[9001:10000, ]
prop.table(table(train$default.payment.next.month))
prop.table(table(test$default.payment.next.month))
# Train the classifier (i.e., decusion tree)
credit_model <- C5.0(train[-24], train$default.payment.next.month)
credit_model
summary(credit_model)
#Evaluate the model
credit_pred <- predict(credit_model, test)
CrossTable(test$default.payment.next.month, credit_pred,prop.chisq = FALSE, prop.c = FALSE, prop.r = FALSE, dnn = c('actual default', 'predicted default'))
confusionMatrix(credit_pred, test$default.payment.next.month, positive = "Y")
|
a6e3d1820e9f78af549205514fe4336d2d17661d
|
f898801224c1f17ba62089b28f3f69c7c525e766
|
/binomial/man/bin_probability.Rd
|
b149aac2493a72bde00e98ddac62ae3ae437c2cc
|
[] |
no_license
|
stat133-sp19/hw-stat133-nadia1212
|
44079944e7b5ab9dffdddbbb3fb82033d2de79a9
|
57ba3ab524660f9d3e8162f1b53a6d030eac6dd6
|
refs/heads/master
| 2020-04-28T12:33:00.104710
| 2019-05-03T19:14:44
| 2019-05-03T19:14:44
| 175,279,474
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 447
|
rd
|
bin_probability.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/functions.R
\name{bin_probability}
\alias{bin_probability}
\title{bin_probability}
\usage{
bin_probability(success, trials, prob)
}
\arguments{
\item{success}{number of success}
\item{trials}{number of trials}
\item{prob}{probability of success}
}
\value{
number of combinations
}
\description{
finds number of combinations in which k success can occur in n trials
}
|
dfa4f2bc4f44c54fec8d7db3de36d71c672e0c72
|
72d9009d19e92b721d5cc0e8f8045e1145921130
|
/spass/inst/testfiles/mlFirstHExp/libFuzzer_mlFirstHExp/libfuzzer_logs/1609981123-inps.R
|
8a1d6fcad90d0447c14a8fc2ea60b5c6155bd97f
|
[] |
no_license
|
akhikolla/TestedPackages-NoIssues
|
be46c49c0836b3f0cf60e247087089868adf7a62
|
eb8d498cc132def615c090941bc172e17fdce267
|
refs/heads/master
| 2023-03-01T09:10:17.227119
| 2021-01-25T19:44:44
| 2021-01-25T19:44:44
| 332,027,727
| 1
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 109
|
r
|
1609981123-inps.R
|
list(kf = 0, tp = 0L, type = 0L, y = numeric(0))
testlist <- list(kf = 0, tp = 0L, type = 0L, y = numeric(0))
|
3c66d0a8d5998ca3f88b983fa45d685f9fcc5974
|
8849921ce5655845b566e5f740fba1a399432fe6
|
/R/CUSUM.R
|
0602c64cba83915bbb70637a23497a89b38b1940
|
[] |
no_license
|
cran/spcadjust
|
3ad36af53c3eb6c0541f81656445177f3ea67b78
|
de5d69322dcb921c7755154e30e5cbd659730f80
|
refs/heads/master
| 2021-01-22T23:53:18.673892
| 2016-09-29T11:37:35
| 2016-09-29T11:37:35
| 17,699,983
| 2
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 5,440
|
r
|
CUSUM.R
|
########################
## Basic CUSUM Charts ##
########################
#' @include main.R model.R CUSUMlib.R
NULL
#' CUSUM Charts
#'
#' Class extending SPCChart with a basic CUSUM charts implementation.
#'
#' The only slot this class contains is the data model. This data
#' model should already incorporate the negative mean for in-control
#' updates that is typical for CUSUM charts.
#'
#' Let \eqn{U_t, t=1,2,\dots} be the updates from the data model. Then
#' the CUSUM chart is given by \eqn{S_0=0} and
#' \deqn{S_t=max(S_{t-1}+U_t,0)}
#'
#' @examples
#' X <- rnorm(1000)
#' chart <- new("SPCCUSUM",model=SPCModelNormal(Delta=1))
#' \dontrun{
#' SPCproperty(data=X,nrep=10,chart=chart,
#' property="calARL",params=list(target=100))
#' SPCproperty(data=X,nrep=10,chart=chart,
#' property="calhitprob",params=list(target=0.05,nsteps=1e3))
#' SPCproperty(data=X,nrep=10,chart=chart,
#' property="ARL",params=list(threshold=3))
#' }
#' SPCproperty(data=X,nrep=10,chart=chart,
#' property="hitprob",params=list(threshold=3,nsteps=1e3))
#' #increase the number of repetitions nrep for real applications.
#'
#' @export
setClass("SPCCUSUM",contains="SPCchart")
#' @describeIn runchart Generic function for running CUSUM
#' charts. Relies on \code{\link{updates}} being implemented for the
#' chart.
#' @export
setMethod("runchart", signature="SPCCUSUM", function(chart,newdata,xi){
R <- cumsum(chart@model$updates(xi=xi, data=newdata))
R - cummin(R)
})
#' @describeIn getq Implements the properties \code{ARL},
#' \code{calARL}, \code{hitprob} and \code{calhitprob}.
#'
#' @export
setMethod("getq", signature="SPCCUSUM",function(chart,property,params){
if (is.null(params[["gridpoints"]])) params$gridpoints=75;
if (grepl("cal",property)&&is.null(params$target))
stop("Argument params contains no element target (needed for calibration).")
if (grepl("hitprob",property)){
if (is.null(params$nsteps))
stop("Argument params does not contain an element nsteps (needed for hitting probabilities).")
else if (params$nsteps<1|round(params$nsteps)!=params$nsteps)
stop("nsteps has to be a positive integer.")
}
if (is.element(property,c("ARL","hitprob"))){
if (is.null(params$threshold))
stop("Argument params does not contain an element threshold.")
else
if (params$threshold<0) stop("Negative threshold.")
}
switch(property,
"calARL"=
list(q= function(P,xi)
log(calibrateARL_Markovapprox(pobs=getcdfupdates(chart,xi=xi, P=P),
ARL=params$target,
gridpoints=params$gridpoints)),
trafo=function(x) exp(x),
lowerconf=TRUE,
format=function(res)
paste("A threshold of ", format(res,digits=4),
" gives an in-control ARL of at least ",
params$target, ".", sep="",collapse="")
),
"ARL"=
list(q= function(P,xi){
as.double(log(ARL_CUSUM_Markovapprox(c=params$threshold,pobs=getcdfupdates(chart,xi=xi, P=P),gridpoints=params$gridpoints)))
},
trafo=function(x) exp(x),
lowerconf=FALSE,
format=function(res)
paste("A threshold of ", params$threshold,
" gives an in-control ARL of at least ",
format(res,digits=4), ".", sep="",collapse="")
),
"hitprob"=
list(q=function(P,xi){
res <- hitprob_CUSUM_Markovapprox(c=params$threshold,pobs=getcdfupdates(chart,xi=xi, P=P),n=params$nsteps,gridpoints=params$gridpoints);
as.double(log(res/(1-res)))
},
trafo=function(x) exp(x)/(1+exp(x)),
lowerconf=TRUE,
format=function(res) paste("A threshold of ", params$threshold,
" gives an in-control false alarm probability of at most ",
format(res,digits=4),
" within ",params$nsteps," steps.",
sep="",collapse="")
),
"calhitprob"=
list(q=function(P,xi)
log(calibratehitprob_Markovapprox(pobs=getcdfupdates(chart,xi=xi, P=P),
hprob=params$target,
n=params$nsteps,
gridpoints=params$gridpoints)),
trafo=function(x) exp(x),
lowerconf=TRUE,
format=function(res) paste("A threshold of ",
format(res,digits=4),
" gives an in-control false alarm probability of at most ",
params$target, " within ",params$nsteps, " steps.",
sep="",collapse="")
),
stop(paste("Property",property,"not implemented."))
)
})
|
6d435a409f3ef5bcabef5dd4a40b2aca3ce00406
|
04eb50424b3fa3e24fff24fbcab6c7b7fcb22f65
|
/scripts/merge_1000.R
|
7d0991d1ef4f4a572f6f926481694db71d5396fc
|
[] |
no_license
|
jonathanperrie/bigham_climber_dump
|
070992955761d26e33009e71a6b420f3c611c33f
|
46884c0c975fb2782f7e70af7751873c1c2071d9
|
refs/heads/main
| 2023-06-21T14:33:00.647978
| 2021-07-29T17:24:49
| 2021-07-29T17:24:49
| 389,523,877
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,955
|
r
|
merge_1000.R
|
library(VariantAnnotation)
setwd("~/Bigham/v2/")
data <- read.table("igsr_samples.tsv",sep="\t",header=T)
x <- scan("intersect_set.txt", what="", sep="\n")
setwd("~/Bigham/v2/1000")
geno1000_list<-read.table("affy_samples.20141118.panel",sep="\t",header=T)
matches_2 <- data[data$Sample.name %in% geno1000_list$sample,]
matches_2 <- matches_2[matches_2$Superpopulation.name=="European Ancestry",]
write.csv(matches_2[c("Sample.name","Sex","Population.code")],"1000geno_meta.csv",quote=F,row.names=FALSE)
genoSum <- function(x) {
suppressWarnings(sum(as.integer(strsplit(x,"/")[[1]])))
}
genoSumOneRow <- function(y){
sapply(y,genoSum)
}
idx<- 1
param <- ScanVcfParam(fixed="ALT", geno=c("GT"))
tab <- TabixFile("ALL.wgs.nhgri_coriell_affy_6.20140825.genotypes_has_ped.vcf.gz", yieldSize=25000)
open(tab)
# first pass to append column names
vcf_yield <- readVcf(tab, param=param)
tmp<-geno(vcf_yield[rownames(vcf_yield) %in% x,colnames(vcf_yield) %in% matches_2$Sample.name])$GT
geno_int <- matrix(nrow=dim(tmp)[1],ncol=dim(tmp)[2])
rownames(geno_int)<-rownames(tmp)
colnames(geno_int)<-colnames(tmp)
for (i in seq(dim(tmp)[1])){
geno_int[i,] <- sapply(tmp[i,],genoSumOneRow)
}
write.table(geno_int,"1000Genomes_intersect.tsv",sep="\t",quote=F,col.names=NA,row.names=TRUE,append=FALSE)
print(idx)
idx <- idx+1
# additional passes to not append column names
while (nrow(vcf_yield <- readVcf(tab, param=param))){
tmp<-geno(vcf_yield[rownames(vcf_yield) %in% x,colnames(vcf_yield) %in% matches_2$Sample.name])$GT
geno_int <- matrix(nrow=dim(tmp)[1],ncol=dim(tmp)[2])
rownames(geno_int)<-rownames(tmp)
colnames(geno_int)<-colnames(tmp)
for (i in seq(dim(tmp)[1])){
geno_int[i,] <- sapply(tmp[i,],genoSumOneRow)
}
write.table(geno_int,"1000Genomes_intersect.tsv",sep="\t",quote=F,col.names=FALSE,row.names=TRUE,append=TRUE)
print(idx)
idx <- idx+1
}
close(tab)
|
50e959bf19c7434a6dadfc261542068458af2b32
|
9ec240c392225a6b9408a1636c7dc6b7d720fd79
|
/packrat/src/backports/backports/man/file.size.Rd
|
0a5914342e5f922492c72c18f7979a62a068119f
|
[] |
no_license
|
wjhopper/PBS-R-Manual
|
6f7709c8eadc9e4f7a163f1790d0bf8d86baa5bf
|
1a2a7bd15a448652acd79f71e9619e36c57fbe7b
|
refs/heads/master
| 2020-05-30T17:47:46.346001
| 2019-07-01T15:53:23
| 2019-07-01T15:53:23
| 189,883,322
| 0
| 1
| null | null | null | null |
UTF-8
|
R
| false
| true
| 795
|
rd
|
file.size.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/file.mode.R, R/file.mtime.R, R/file.size.R
\name{file.mode}
\alias{file.mode}
\alias{file.mtime}
\alias{file.size}
\title{Backports of wrappers around \code{file.info} for R < 3.2.0}
\usage{
file.mode(...)
file.mtime(...)
file.size(...)
}
\description{
See the original description in \code{base::file.size}.
}
\examples{
# get functions from namespace instead of possibly getting
# implementations shipped with recent R versions:
bp_file.size = getFromNamespace("file.size", "backports")
bp_file.mode = getFromNamespace("file.size", "backports")
bp_file.mtime = getFromNamespace("file.size", "backports")
fn = file.path(R.home(), "COPYING")
bp_file.size(fn)
bp_file.mode(fn)
bp_file.size(fn)
}
\keyword{internal}
|
9139adf60c030d3435badf6065649dfe0cc263b7
|
79aa103b7b35ae807444be74805e42ed4e57acc5
|
/R/ggscatter.R
|
ecd70c1d37ca72f981baa5fe34adf14dd4bbd870
|
[
"MIT"
] |
permissive
|
UBC-MDS/ggexpress
|
47ac0252e35ae6f7ee0cc52743bd303b45219d09
|
76e9692c62054a71534d8da1c98f03c424e3a4cc
|
refs/heads/master
| 2021-01-15T06:03:34.739464
| 2020-03-26T22:45:36
| 2020-03-26T22:45:36
| 242,897,067
| 2
| 1
|
NOASSERTION
| 2020-03-26T22:45:38
| 2020-02-25T03:06:09
|
R
|
UTF-8
|
R
| false
| false
| 1,872
|
r
|
ggscatter.R
|
library(dplyr)
library(ggplot2)
#' Create a scatterplot and calculate correlation values for two numerical variables
#'
#' Creates a ggplot scatterplot object containing two numerical variables. Arguments in the
#' function will control whether correlational values and log transformations are calculated for the input data.
#'
#' @param df Dataframe to plot
#' @param xval x-var Column name used as the x-axis variable
#' @param yval y-var Column name used as the y-axis variable
#' @param x_transform Determines whether the x-axis undergoes a natural log transformation
#' @param y_transform Determines whether the y-axis undergoes a natural log transformation
#'
#' @return ggplot Object
#' @export
#'
#' @examples
#' scatter_express(iris, Sepal.Width, Sepal.Length)
scatter_express <- function(df, xval = NA, yval = NA, x_transform = FALSE, y_transform = FALSE){
corr_df <- dplyr::select(df, {{xval}}, {{yval}})
if (is.numeric(corr_df[[1]]) == FALSE) stop("Your x-variable must be numeric")
if (is.numeric(corr_df[[2]]) == FALSE) stop("Your y-variable must be numeric")
corr_val <- round(stats::cor(corr_df[[1]], corr_df[[2]]), 2)
scatter <- ggplot2::ggplot(corr_df, ggplot2::aes(x = {{xval}}, y = {{yval}})) +
ggplot2::geom_point(color = "blue")
if (x_transform == TRUE) {
scatter <- scatter + ggplot2::scale_x_continuous(trans = "log2")
}
if (y_transform == TRUE) {
scatter <- scatter + ggplot2::scale_y_continuous(trans = "log2")
}
scatter <- scatter + ggplot2::labs(title = paste(rlang::get_expr(scatter$mapping$x),
" vs ", rlang::get_expr(scatter$mapping$y),
"(Pearson Correlation: ",
corr_val,
")"))
scatter
}
|
74cdbc5eec50d85ca2197607cafb639dead4b79f
|
d11fadd13f28e7223c37a8395661cc650159c817
|
/R/urbn_geofacet.R
|
7d895f16d4526fdae40acf784efe2f54c15d263c
|
[] |
no_license
|
scoultersdcoe/urbnthemes
|
4f5a7dd3a76bc68a1b2709fc866e737503a49062
|
8d88a618c8bf50a4c8c4a80eef3bfd5f5f4e9200
|
refs/heads/master
| 2023-08-01T02:59:26.110414
| 2021-09-03T20:58:11
| 2021-09-03T20:58:11
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 284
|
r
|
urbn_geofacet.R
|
#' Urban Institute geofacet template
#'
#' @format Data frame with columns
#' \describe{
#' \item{row}{Row in geofacet}
#' \item{col}{Column in geofacet}
#' \item{state_abbv}{State abbreviation}
#' \item{state_name}{State name}
#' }
"urbn_geofacet"
#' @importFrom tibble tibble
NULL
|
5f8d01ae57a3d97ee718f3c859ea198761d04b46
|
95f82fae345c99b0ac48f7080306617d8af109da
|
/run_analysis.R
|
f9f3a464ea2681b8df13035dcd6acdec40e8c837
|
[] |
no_license
|
PavelMAA/Getting-and-Cleaning-Data-Course-Project
|
681ac6a7e06a68081f5a05012bbfdbb4ebf10a26
|
a965fcf2ee7b4b69b420fc4e7090f47760ba20e3
|
refs/heads/master
| 2023-06-28T02:00:42.938567
| 2021-08-06T00:58:28
| 2021-08-06T00:58:28
| 393,202,736
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 2,974
|
r
|
run_analysis.R
|
rm(list = ls())
##install library
install.packages("reshape2")
install.packages("dplyr")
library(reshape2)
library(dplyr)
## define directory
setwd("C:/Coursera/GettingData")
#Get data(data> UCI HAR Dataset)
if(!file.exists("./data")){dir.create("./data")}
fileUrl <- "https://d396qusza40orc.cloudfront.net/getdata%2Fprojectfiles%2FUCI%20HAR%20Dataset.zip"
download.file(fileUrl,destfile="./data/Dataset.zip",method="curl")
#unzip data in the folder (data> UCI HAR Dataset)
unzip(zipfile="./data/Dataset.zip",exdir="./data")
#define the path
p_rf <- file.path("./data" , "UCI HAR Dataset")
f<-list.files(p_rf, recursive=TRUE)
f
#Read the data from the files
dActivityTest <- read.table(file.path(p_rf, "test" , "Y_test.txt" ),header = FALSE)
dActivityTrain <- read.table(file.path(p_rf, "train", "Y_train.txt"),header = FALSE)
dSubjectTrain <- read.table(file.path(p_rf, "train", "subject_train.txt"),header = FALSE)
dSubjectTest <- read.table(file.path(p_rf, "test" , "subject_test.txt"),header = FALSE)
dFeaturesTest <- read.table(file.path(p_rf, "test" , "X_test.txt" ),header = FALSE)
dFeaturesTrain <- read.table(file.path(p_rf, "train", "X_train.txt"),header = FALSE)
##Check the properties of the variable
str(dActivityTest)
str(dActivityTrain)
str(dSubjectTest)
str(dSubjectTrain)
str(dFeaturesTest)
str(dFeaturesTrain)
#Merges the training and the test sets to create one data set
dSubject <- rbind(dSubjectTrain, dSubjectTest)
dActivity<- rbind(dActivityTrain, dActivityTest)
dFeatures<- rbind(dFeaturesTrain, dFeaturesTest)
# Tidy the variable name
names(dSubject)<-c("subject")
names(dActivity)<- c("activity")
# read the feature.txt
dFeaturesNames <- read.table(file.path(p_rf, "features.txt"),head=FALSE)
names(dFeatures)<- dFeaturesNames$V2
# Merge the data set
dCombine <- cbind(dSubject, dActivity)
D <- cbind(dFeatures, dCombine)
# Calculate the mean and standard deviation
subdFeaturesNames<-dFeaturesNames$V2[grep("mean\\(\\)|std\\(\\)", dFeaturesNames$V2)]
# Add the "subject", "activity" with the subdFeaturesNames
selectedNames<-c(as.character(subdFeaturesNames), "subject", "activity" )
# Create again data set base on the selectedNames
D<-subset(D,select=selectedNames)
str(D)
#Uses descriptive activity names to name of the activities in the data set
activityLabels <- read.table(file.path(p_rf, "activity_labels.txt"),header = FALSE)
head(D$activity,30)
#Appropriately labels the data set with descriptive variable names
names(D)<-gsub("^t", "time", names(D))
names(D)<-gsub("^f", "frequency", names(D))
names(D)<-gsub("Acc", "Accelerometer", names(D))
names(D)<-gsub("Gyro", "Gyroscope", names(D))
names(D)<-gsub("Mag", "Magnitude", names(D))
names(D)<-gsub("BodyBody", "Body", names(D))
names(D)
#Creates a second,independent tidy data set and it output
library(plyr);
D2<-aggregate(. ~subject + activity, D, mean)
D2<-D2[order(D2$subject,D2$activity),]
write.table(D2, file = "tidydataset.txt",row.name=FALSE)
|
441abfb264f145b9b801ab942b3c3de5f2183669
|
f53c62df2e61aa215870b40ce90265bf97705960
|
/RNA_FISH/AC16_Tp53_Results.R
|
87667bd8ee5e3c168aa519f5f95d7b676ccd7404
|
[
"MIT"
] |
permissive
|
hjf34/Cold
|
058b3ac970a20ef7902909d3d910261a784b6dbc
|
32c70926e4646e8db62f0e6752efe73dc8a19bc1
|
refs/heads/master
| 2023-02-23T03:06:37.524385
| 2021-01-19T15:32:03
| 2021-01-19T15:32:03
| 272,039,850
| 1
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 3,911
|
r
|
AC16_Tp53_Results.R
|
setwd("/path/to/Results.csv/directory/")
lf = list.files()
lf1 = lf[grep("Results.csv", lf)]
filenamestem = as.vector(sapply(lf1, function(n) strsplit(n, ".dv")[[1]][1]))
Nr1d1NuclearDots = rep(NA, length(lf1))
Nr1d1CytoplasmicDots = rep(NA, length(lf1))
NuclearArea = rep(NA, length(lf1))
WholeCellArea = rep(NA, length(lf1))
for(a1 in 1:length(lf1)){
csv1 = read.csv((lf1[a1]), stringsAsFactors = F)
NuclearArea[a1] = csv1$Area[1]
WholeCellArea[a1] = csv1$Area[2]
Nr1d1NuclearDots[a1] = csv1$Count[3]
Nr1d1CytoplasmicDots[a1] = csv1$Count[4]
}
condition_repeat = as.vector(sapply(lf1, function(n) strsplit(n, "_Poor")[[1]][1]))
condition = as.vector(sapply(lf1, function(n) strsplit(n, "_Rep")[[1]][1]))
df = data.frame(condition_repeat, condition, Nr1d1NuclearDots,Nr1d1CytoplasmicDots,NuclearArea,WholeCellArea)
df$CytoplasmicArea = df$WholeCellArea - df$NuclearArea
df$cba = df$Nr1d1CytoplasmicDots/df$CytoplasmicArea
df$nba = df$Nr1d1NuclearDots/df$NuclearArea
df1 = df
row.names(df1) = filenamestem
colnames(df1)[c(8,9)] = c("DotsPerCyt","DotsPerNuc")
repsimaged = data.frame(images=c(table(df1$condition_repeat)))
write.csv(repsimaged, file="AC16_Tp53_RepsImaged.csv")
write.csv(df1, file="AC16_Tp53_Dots_Per_Image.csv")
##########################################################################
##########################################################################
dba = df[,c("condition","cba","nba")]
dba$condition = as.vector(dba$condition)
dba1 = dba
dba1$condition1 = factor(dba1$condition, levels=names(table(dba1$condition))[c(3,1,2)])
###########################################################################
########PER IMAGE PLOT AND STATS
mx1 = aggregate(dba1[,2:3], by=list(dba1$condition1), function(n) mean(n))
sx1 = aggregate(dba1[,2:3], by=list(dba1$condition1), function(n) sd(n)/sqrt(length(n)))
m1 = mx1[,3]; m2 = mx1[,2]
sem1 = sx1[,3]; sem2 = sx1[,2]
#################################
#################################
barplotter = function(means,sems,colours){
mx = means
smx = sems
par(mar = c(1,5,1,1))
mp = barplot(mx, cex.axis=1.5,col=colours,
ylim = c(0,1.2*max(mx+smx, na.rm=T)),xaxt="n", las=1, main = "", ylab = "")
subseg1 = as.numeric(mx) -as.numeric(smx)
subseg2 = as.numeric(mx) +as.numeric(smx)
subseg1[subseg1 < 0] = 0
segments(mp, subseg1, mp, subseg2, lwd=2)
segments(mp-0.2, subseg1, mp+0.2, subseg1, lwd=2)
segments(mp-0.2, subseg2, mp+0.2, subseg2, lwd=2)
box(lwd = 2)
}
colours1 = c("gold","dodgerblue3","purple")
#########################NUCLEAR
png("AC16_Tp53NucPerImageMeanSEM.png", height = 3, width = 2.5, units = "in", res = 600)
barplotter(m1,sem1,colours1)
dev.off()
#########################CYTOPLASMIC
png("AC16_Tp53CytPerImageMeanSEM.png", height = 3, width = 2.5, units = "in", res = 600)
barplotter(m2,sem2,colours1)
dev.off()
###########################################################################
###########################################################################
###########################################################################
summary(aov(cba~condition1, data=dba1))
cTHSD = TukeyHSD(aov(cba~condition1, data=dba1))$condition1
summary(aov(nba~condition1, data=dba1))
nTHSD = TukeyHSD(aov(nba~condition1, data=dba1))$condition1
aovdata = rbind(data.frame(summary(aov(nba~condition1, data=dba1))[[1]]),data.frame(summary(aov(cba~condition1, data=dba1))[[1]]))
row.names(aovdata) = c("Nuc_Condition","Nuc_Residuals","Cyt_Condition","Cyt_Residuals")
THSD = rbind(nTHSD,cTHSD)
row.names(THSD) = c(paste("Nuc_",row.names(THSD)[1:3], sep=""),paste("Cyt_",row.names(THSD)[4:6], sep=""))
write.csv(THSD, file="AC16_Tp53_TukeyHSDdata.csv")
write.csv(aovdata, file="AC16_Tp53_ANOVAdata.csv")
#############################################################################
#############################################################################
|
44b23c880aa8e07ec657d656bc673732b05d359c
|
58ec8c9b97ea1bd69aed7d55c994f026aad420a2
|
/man/myLinearRegression.Rd
|
9ad23a596fc9d193f55a1314881351bfa4377c23
|
[] |
no_license
|
msalmon7/myLinearRegressionPackage
|
6a888f366a806a30809e86e9549b39d9dee53cdc
|
de07ff8619f8b9483515380a022df738b9c43da6
|
refs/heads/master
| 2022-06-12T07:36:05.779333
| 2020-05-01T15:39:08
| 2020-05-01T15:39:08
| 260,494,542
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 1,716
|
rd
|
myLinearRegression.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/myLinearRegression.R
\name{myLinearRegression}
\alias{myLinearRegression}
\title{Perform linear regression and create scatterplots of each pair of covariates.}
\usage{
myLinearRegression(y = myData[, 1], x = myData[, 2:4], sub = c(1:20))
}
\arguments{
\item{y}{A vector of outcomes.}
\item{x}{A matrix of covariates.}
\item{sub}{A list of subjects (i.e. a set of integers corresponding to rows in x)}
}
\value{
The coefficients and p-values of a linear regression performed on \code{y} subject to \code{x} and a scatterplot matrix of each pair of covariates.
}
\description{
This function takes inputs of a vector as the dependent variable,
a matrix as a set of independent variables, and a list of subjects
from these sets to perform linear regression. The function
outputs the coefficients and p-values from the regression
as well as a scatterplot matrix barring that there are more than 5
covariates, in which case a warning is given.
}
\examples{
# You can reference the data any way you would like. For example, the
# data set myData is running a linear regression of column "Y" and columns
# "X1", "X2", and "X3." This data set has 100 rows, but we're using the sub
# function to specify that we only want to look at the first 30.
myLinearRegression(y = myData[, "Y"], x = myData[ ,c("X1", "X2", "X3")], sub = c(1:30))
# Similarly, you can create your own vector to perform linear regression with.
# If 5 or more columns are selected as covariates, as is the case here, the
# function will not output any scatterplots. This only looks at rows 5 through 20.
myLinearRegression(y = c(1:50), x = myData[ ,2:6], sub = c(5:20))
}
|
621cb6f6284d288722ea7ba040069c127d0a5f22
|
10af39cbdd712d0bd786232d41089b71005788aa
|
/data_raw/results_exp_ols_mc.mvn_ci.R
|
6f0099a4c0792962599f3c4d218eb4dd738c63f1
|
[
"MIT"
] |
permissive
|
jeksterslabds/jeksterslabRmedsimple
|
1f914691b75c44f14ea6701bdbec776908997da5
|
4a14bc41892bfa670ebc56f691262475986d48fb
|
refs/heads/master
| 2022-12-30T23:59:12.475315
| 2020-10-16T05:48:55
| 2020-10-16T05:48:55
| 287,948,188
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,123
|
r
|
results_exp_ols_mc.mvn_ci.R
|
#' ---
#' title: "Data: Simple Mediation Model - Exponential X lambda = 1 - Complete Data - Monte Carlo Method Confidence Intervals with Ordinary Least Squares Parameter Estimates and Standard Errors"
#' author: "Ivan Jacob Agaloos Pesigan"
#' date: "`r Sys.Date()`"
#' output:
#' rmarkdown::html_vignette:
#' toc: true
#' vignette: >
#' %\VignetteIndexEntry{Data: Simple Mediation Model - Exponential X lambda = 1 - Complete Data - Monte Carlo Method Confidence Intervals with Ordinary Least Squares Parameter Estimates and Standard Errors}
#' %\VignetteEngine{knitr::rmarkdown}
#' %\VignetteEncoding{UTF-8}
#' ---
#'
#+ data
results_exp_ols_mc.mvn_ci <- readRDS("summary_medsimple_exp_ols_mc.mvn_pcci.Rds")
# subset
# results_exp_ols_mc.mvn_ci <- results_exp_ols_mc.mvn_ci[which(results_exp_ols_mc.mvn_ci$taskid < 145 | results_exp_ols_mc.mvn_ci$taskid > 153), ]
# results_exp_ols_mc.mvn_ci <- results_exp_ols_mc.mvn_ci[which(results_exp_ols_mc.mvn_ci$n > 20), ]
head(results_exp_ols_mc.mvn_ci)
str(results_exp_ols_mc.mvn_ci)
#'
#+ usedata
usethis::use_data(
results_exp_ols_mc.mvn_ci,
overwrite = TRUE
)
|
49be68b4186cc4bf35bf4254668cd9f1872b2a94
|
7d5d8492c2d88b88bdc57e3c32db038a7e7e7924
|
/PhD/0007-crop-modelling/scripts/cmip5/06.bc_rain-functions.R
|
5d07755a84e26bc7b9652dddcd94f375b24e3cb3
|
[] |
no_license
|
CIAT-DAPA/dapa-climate-change
|
80ab6318d660a010efcd4ad942664c57431c8cce
|
2480332e9d61a862fe5aeacf6f82ef0a1febe8d4
|
refs/heads/master
| 2023-08-17T04:14:49.626909
| 2023-08-15T00:39:58
| 2023-08-15T00:39:58
| 39,960,256
| 15
| 17
| null | null | null | null |
UTF-8
|
R
| false
| false
| 13,291
|
r
|
06.bc_rain-functions.R
|
#Julian Ramirez-Villegas
#UoL / CIAT / CCAFS
#Oct 2012
#final wrap
wrap_bc_wthmkr <- function(k) {
source(paste(src.dir,"/0007-crop-modelling/scripts/cmip5/06.bc_rain-functions.R",sep=""))
lmts <- bc_rain_wrapper(k) # bias correct the data
ctrf <- make_bc_wth_wrapper(k) # generate the wth files
}
#make bias corrected weather files for GLAM runs
make_bc_wth_wrapper <- function(i) {
library(raster)
#source functions of interest
source(paste(src.dir,"/0006-weather-data/scripts/GHCND-GSOD-functions.R",sep=""))
source(paste(src.dir,"/0008-CMIP5/scripts/CMIP5-functions.R",sep=""))
source(paste(src.dir,"/0007-crop-modelling/scripts/cmip5/06.bc_rain-functions.R",sep=""))
source(paste(src.dir,"/0007-crop-modelling/scripts/cmip5/01.make_wth-functions.R",sep=""))
source(paste(src.dir,"/0007-crop-modelling/scripts/glam/glam-make_wth.R",sep=""))
sce <- paste(all_proc$GCM[i])
gcm <- unlist(strsplit(sce,"_ENS_",fixed=T))[1]
ens <- unlist(strsplit(sce,"_ENS_",fixed=T))[2]
loc <- all_proc$LOC[i] #635
thisLeap <- paste(gcmChars$has_leap[which(gcmChars$GCM == gcm)][1])
#directories where the uncorrected data is
inDir_his <- paste(hisDir,"/",gcm,"/",ens,sep="") #folder with raw gridded data
inDir_rcp <- paste(rcpDir,"/",gcm,"/",ens,sep="") #folder with raw gridded data
#directories where the bias corrected data is
oDir_his_bc <- paste(bcDir_his,"/",gcm,"/",ens,sep="") #folder with bc gridded data
oDir_rcp_bc <- paste(bcDir_rcp,"/",gcm,"/",ens,sep="") #folder with bc gridded data
#directories where check files are
checkDir <- paste(wthDirBc_his,"/_process",sep="")
if (!file.exists(checkDir)) {dir.create(checkDir)}
checkFil_his <- paste(checkDir,"/",sce,"_loc-",loc,".proc",sep="")
#historical period
if (!file.exists(checkFil_his)) {
#copy all other data from the uncorrected output, and then remove it
for (cvn in c("rsds","tasmax","tasmin")) {
codir <- paste(oDir_his_bc,"/",cvn,sep="")
if (!file.exists(codir)) {dir.create(codir)}
ff <- file.copy(from=paste(inDir_his,"/",cvn,"/cell-",loc,".csv",sep=""),to=codir)
}
#create the daily data files for historical
outfol_his <- write_cmip5_loc(all_locs=cells,gridcell=loc,scen=sce,
year_i=1966,year_f=1993,wleap=thisLeap,
out_wth_dir=wthDirBc_his,fut_wth_dir=oDir_his_bc,
sow_date_dir=sowDir)
#remove extra files
for (cvn in c("rsds","tasmax","tasmin")) {
codir <- paste(oDir_his_bc,"/",cvn,sep="")
ff <- file.remove(from=paste(codir,"/cell-",loc,".csv",sep=""),to=codir)
#system(paste("rm -rf ",codir,sep=""))
}
ff <- file(checkFil_his,"w")
cat("Processed on",date(),"\n",file=ff)
close(ff)
}
#directories where check files are
checkDir <- paste(wthDirBc_rcp,"/_process",sep="")
if (!file.exists(checkDir)) {dir.create(checkDir)}
checkFil_rcp <- paste(checkDir,"/",sce,"_loc-",loc,".proc",sep="")
if (!file.exists(checkFil_rcp)) {
#copy all other data from the uncorrected output, and then remove it
for (cvn in c("rsds","tasmax","tasmin")) {
codir <- paste(oDir_rcp_bc,"/",cvn,sep="")
if (!file.exists(codir)) {dir.create(codir)}
ff <- file.copy(from=paste(inDir_rcp,"/",cvn,"/cell-",loc,".csv",sep=""),to=codir)
}
outfol_rcp <- write_cmip5_loc(all_locs=cells,gridcell=loc,scen=sce,
year_i=2021,year_f=2049,wleap=thisLeap,
out_wth_dir=wthDirBc_rcp,fut_wth_dir=oDir_rcp_bc,
sow_date_dir=sowDir)
#remove extra files
for (cvn in c("rsds","tasmax","tasmin")) {
codir <- paste(oDir_rcp_bc,"/",cvn,sep="")
ff <- file.remove(from=paste(codir,"/cell-",loc,".csv",sep=""),to=codir)
#system(paste("rm -rf ",codir,sep=""))
}
ff <- file(checkFil_rcp,"w")
cat("Processed on",date(),"\n",file=ff)
close(ff)
}
return(list(HIS=checkFil_his,RCP=checkFil_rcp))
}
#bias correction wrapper function
bc_rain_wrapper <- function(i) {
library(raster)
#source functions of interest
source(paste(src.dir,"/0006-weather-data/scripts/GHCND-GSOD-functions.R",sep=""))
source(paste(src.dir,"/0008-CMIP5/scripts/CMIP5-functions.R",sep=""))
source(paste(src.dir,"/0007-crop-modelling/scripts/cmip5/06.bc_rain-functions.R",sep=""))
source(paste(src.dir,"/0007-crop-modelling/scripts/cmip5/01.make_wth-functions.R",sep=""))
source(paste(src.dir,"/0007-crop-modelling/scripts/glam/glam-make_wth.R",sep=""))
sce <- paste(all_proc$GCM[i])
gcm <- unlist(strsplit(sce,"_ENS_",fixed=T))[1]
ens <- unlist(strsplit(sce,"_ENS_",fixed=T))[2]
loc <- all_proc$LOC[i] #635
thisLeap <- paste(gcmChars$has_leap[which(gcmChars$GCM == gcm)][1])
cat("processing",sce," / and loc = ",loc,"\n")
#specific gcm directories
oDir_his <- paste(bcDir_his,"/",gcm,"/",ens,"/",vn_gcm,sep="")
oDir_rcp <- paste(bcDir_rcp,"/",gcm,"/",ens,"/",vn_gcm,sep="")
if (!file.exists(oDir_his)) {dir.create(oDir_his,recursive=T)}
if (!file.exists(oDir_rcp)) {dir.create(oDir_rcp,recursive=T)}
#############################################################################
# bias correcting the GCM output precipitation
#############################################################################
if (!file.exists(paste(oDir_his,"/fit_cell-",loc,".csv",sep=""))) {
#read in observed and gcm data
obs_data <- read.csv(paste(obsDir,"/",vn,"/cell-",loc,".csv",sep=""))
his_data <- read.csv(paste(hisDir,"/",gcm,"/",ens,"/",vn_gcm,"/cell-",loc,".csv",sep=""))
rcp_data <- read.csv(paste(rcpDir,"/",gcm,"/",ens,"/",vn_gcm,"/cell-",loc,".csv",sep=""))
#separate months into individual series
obs_list <- mk_mth_list(obs_data,"createDateGrid",dg_args=NA,yi_h,yf_h)
his_list <- mk_mth_list(his_data,"createDateGridCMIP5",dg_args=thisLeap,yi_h,yf_h)
rcp_list <- mk_mth_list(rcp_data,"createDateGridCMIP5",dg_args=thisLeap,yi_f,yf_f)
#calculate loci metrics
loci_mets <- loci_cal(obs_list,his_list,wdt_obs=1,iter_step=1e-4)
#calculate bias corrected data based on historical data
his_list <- loci_correct(his_list,loci_mets)
#putting all the series together again into a matrix
his_data_bc <- remake_daily(his_list,his_data,thisLeap,yi_h,yf_h)
#here bias correct the future climates and plot the pdfs together
rcp_list <- loci_correct(rcp_list,loci_mets)
rcp_data_bc <- remake_daily(rcp_list,rcp_data,thisLeap,yi_f,yf_f)
#here write the data
write.csv(his_data_bc,paste(oDir_his,"/cell-",loc,".csv",sep=""),quote=T,row.names=F)
write.csv(rcp_data_bc,paste(oDir_rcp,"/cell-",loc,".csv",sep=""),quote=T,row.names=F)
write.csv(loci_mets,paste(oDir_his,"/fit_cell-",loc,".csv",sep=""),quote=T,row.names=F)
} else {
loci_mets <- read.csv(paste(oDir_his,"/fit_cell-",loc,".csv",sep=""))
}
return(loci_mets)
}
#calculate total rainfall and number of rainy days for the whole time series
calc_metrics <- function(all_data,dg_fun="createDateGrid",dg_args=NA,yi,yf) {
cat("calculating rainfall and rain days\n")
odat_all <- data.frame()
for (yr in yi:yf) {
yr_data <- as.numeric(all_data[which(all_data$YEAR == yr),2:ncol(all_data)])
if (!is.na(dg_args)) {
dg <- do.call(dg_fun,list(yr,dg_args))
} else {
dg <- do.call(dg_fun,list(yr))
}
dg$MTH <- as.numeric(substr(dg$MTH.DAY,2,3))
dg$VALUE <- yr_data[1:nrow(dg)]
pr <- as.numeric(by(dg$VALUE,dg$MTH,FUN=sum))
rd <- as.numeric(by(dg$VALUE,dg$MTH,FUN=function(x) {length(which(x>=1))}))
pr_jja <- sum(pr[6:8])
rd_jja <- sum(rd[6:8])
pr_ann <- sum(pr)
rd_ann <- sum(rd)
odat <- data.frame(YEAR=yr,MTH=c(1:12,"JJA","ANN"),PR=c(pr,pr_jja,pr_ann),RD=c(rd,rd_jja,rd_ann))
odat_all <- rbind(odat_all,odat)
}
return(odat_all)
}
#re construct a daily data.frame using a dummy one and the corrected data
remake_daily <- function(his_list,bc_data,wleap,yi,yf) {
cat("remaking daily data.frame\n")
bc_data[,2:ncol(bc_data)] <- NA
for (mth in 1:12) {
#cat("calculating month",mth,"\n")
#looping through years
out_df <- data.frame()
for (yr in yi:yf) {
dg <- createDateGridCMIP5(yr,wleap)
dg$MTH <- as.numeric(substr(dg$MTH.DAY,2,3))
jdi <- min(dg$JD[which(dg$MTH == mth)])+1 #+1 to account that 1st column is year
jdf <- max(dg$JD[which(dg$MTH == mth)])+1
data_yr <- his_list[[paste("MTH.",mth,sep="")]]
data_yr <- data_yr$VALUE_BC[which(data_yr$YEAR==yr)]
bc_data[which(bc_data$YEAR==yr),jdi:jdf] <- data_yr
}
}
return(bc_data)
}
#correct a given time series based on pre-fitted parameters
loci_correct <- function(his_data,loci_mets) {
cat("correcting the data\n")
#now apply the correction to the time series
for (mth in 1:12) {
s <- loci_mets$S[which(loci_mets$MTH==mth)]
wdt_obs <- loci_mets$WDT_OBS[which(loci_mets$MTH==mth)]
wdt_mod <- loci_mets$WDT_MOD[which(loci_mets$MTH==mth)]
nwd_obs <- loci_mets$NWD_OBS[which(loci_mets$MTH==mth)]
his_data[[paste("MTH.",mth,sep="")]]$VALUE_BC <- sapply(his_data[[paste("MTH.",mth,sep="")]]$VALUE,FUN=function(x) {y<-max(c(0,s*(x-wdt_mod)+wdt_obs));if (nwd_obs == 0) {y<-0}; return(y)})
}
return(his_data)
}
#calibrate loci for all momths
loci_cal <- function(obs_list,his_list,wdt_obs=1,iter_step=0.0001) {
cat("calculating loci metrics\n")
out_mets <- data.frame()
for (mth in 1:12) {
ts_obs <- obs_list[[paste("MTH.",mth,sep="")]]
ts_mod <- his_list[[paste("MTH.",mth,sep="")]]
mth_mx <- loci_cal_mth(ts_obs,ts_mod,wdt_obs=wdt_obs,iter_step=iter_step)
mth_mx <- cbind(MTH=mth,mth_mx)
out_mets <- rbind(out_mets,mth_mx)
}
return(out_mets)
}
#calibrate loci for a month
loci_cal_mth <- function(ts_obs,ts_mod,wdt_obs=1,iter_step=0.0001) {
#calculate number of wet days and average rain in wet days
wdays_obs <- which(ts_obs$VALUE>=wdt_obs)
nwd_obs <- length(wdays_obs)
if (nwd_obs==0) {
wet_obs <- 0
} else {
wet_obs <- mean(ts_obs$VALUE[wdays_obs],na.rm=T)
}
#find wet-day threshold for GCM
wdt_mod <- find_wdt(ts_mod$VALUE,nwd_obs,iter_step=iter_step)
wdays_mod <- which(ts_mod$VALUE>=wdt_mod)
nwd_mod <- length(wdays_mod)
if (nwd_mod==0) {
wet_mod <- 0
} else {
wet_mod <- mean(ts_mod$VALUE[wdays_mod],na.rm=T)
}
#calculate s correction factor
s <- (wet_obs-wdt_obs)/(wet_mod-wdt_mod)
#put everything into a matrix
out_mx <- data.frame(WDT_OBS=wdt_obs,NWD_OBS=nwd_obs,WET_OBS=wet_obs,
WDT_MOD=wdt_mod,NWD_MOD=nwd_mod,WET_MOD=wet_mod,S=s)
return(out_mx)
}
#find the GCM wet-day threshold that matches the observed ones
find_wdt <- function(values,nwd_obs,iter_step=1e-4) {
nwd_mod <- length(values)+1 #initialise
wdt_mod <- iter_step*-1 #initialise
if (length(which(values>=0)) < nwd_obs) {
wdt_mod <- 0
} else {
nwitr <- 0
while (nwd_mod > nwd_obs) {
wdt_mod <- wdt_mod+iter_step
nwd_mod <- length(which(values>=wdt_mod))
nxt_nwd <- length(which(values>=(wdt_mod+iter_step)))
#if the next value exceeds the value i'm looking for
if (nxt_nwd < nwd_obs) {
new_istep <- iter_step
niter <- 0
#until a value of iter_step is found so that it does not exceed
#the value i'm looking for
while (nxt_nwd < nwd_obs) {
new_istep <- new_istep*0.1
nxt_nwd <- length(which(values>=(wdt_mod+new_istep)))
niter <- niter+1
#if (niter == 100) {nxt_nwd <- nwd_obs}
}
iter_step <- new_istep
}
nwitr <- nwitr+1
#truncation
if (nwitr==100000) {nwd_mod <- nwd_obs}
}
}
return(wdt_mod)
}
#### make a monthly list from the yearly matrices
mk_mth_list <- function(all_data,dg_fun="createDateGrid",dg_args=NA,yi,yf) {
cat("making monthly list\n")
out_all <- list()
for (mth in 1:12) {
#cat("calculating month",mth,"\n")
#looping through years
out_df <- data.frame()
for (yr in yi:yf) {
if (!is.na(dg_args)) {
dg <- do.call(dg_fun,list(yr,dg_args))
} else {
dg <- do.call(dg_fun,list(yr))
}
#get days of interest (julian days of month mth)
dg$MTH <- as.numeric(substr(dg$MTH.DAY,2,3))
jdi <- min(dg$JD[which(dg$MTH == mth)])+1 #+1 to account that 1st column is year
jdf <- max(dg$JD[which(dg$MTH == mth)])+1
#get the data from the matrix
data_yr <- all_data[which(all_data$YEAR==yr),]
data_yr <- as.numeric(data_yr[,jdi:jdf])
tmp_df <- data.frame(YEAR=yr,VALUE=data_yr)
out_df <- rbind(out_df,tmp_df)
}
out_all[[paste("MTH.",mth,sep="")]] <- out_df
}
return(out_all)
}
|
6d2a5deb4e11cad50cf670d0ba58a0623088554a
|
09b581d3c65d6c9687684aa5e538e3b650f130d2
|
/Cross-validation.R
|
77257f341da16f357b6ca3d518d576c2eaf55abf
|
[] |
no_license
|
seth127/statToolkit
|
a9fe2886341eb609def4cf9497fdcb4413a9f3d0
|
e220cd3d760cadd6bce663e92eb72f5fa9a82625
|
refs/heads/master
| 2021-01-11T04:40:10.916944
| 2016-10-26T04:36:00
| 2016-10-26T04:36:00
| 71,141,777
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 8,273
|
r
|
Cross-validation.R
|
setwd("~/Documents/DSI/notes/2-STAT-6021/team assignments")
train <- read.csv("teamassign05train.csv", header=T, stringsAsFactors = F)
## cross validate with the lm function
cv.lm <- function(vars, train, k=5) { ## vars should be a character vector of variable names/combinations
# the function to do the cross-validation
theCV <- function(var, train, k, seed) {
# create formula
form = paste('y ~', var)
#make column for preds
train$preds <- numeric(nrow(train))
#randomize the indexes
nums <- sample(row.names(train), nrow(train))
#split the indexes into k groups
nv <- split(nums, cut(seq_along(nums), k, labels = FALSE))
#subset the training data into k folds
trainlist <- list()
for (i in 1:k) {
trainlist[[i]] <- train[nv[[i]], ]
}
#trainlist
#run on each fold
for (i in 1:k) {
ftrainlist <- trainlist[-i]
ftrain <- ftrainlist[[1]]
for (j in 2:length(ftrainlist)) {
ftrain <- rbind(ftrain, ftrainlist[[j]])
}
mod <- lm(as.formula(paste(form,' - preds')), data = ftrain) ### the model
trainlist[[i]]$preds <- predict(mod, newdata = trainlist[[i]])
}
#reassemble
cvdata <- ftrainlist[[1]]
for (j in 2:length(trainlist)) {
cvdata <- rbind(cvdata, trainlist[[j]])
}
# cross-validated test set MSE
###degfree <- nrow(cvdata) - ncol(subset(cvdata, select = -c(y, preds))) ##just use n?
MSE <- sum((cvdata$y-cvdata$preds)^2) / nrow(cvdata)
##training set stats
m <- lm(as.formula(paste(form,' - preds')), data = train)
# adjusted R-squared
aR2 <- summary(m)$adj.r.squared
# p-value from F-stat
lmp <- function (modelobject) {
if (class(modelobject) != "lm") stop("Not an object of class 'lm' ")
f <- summary(modelobject)$fstatistic
p <- pf(f[1],f[2],f[3],lower.tail=F)
attributes(p) <- NULL
return(p)
}
p <- lmp(m)
list(form, MSE, aR2, p)
}
#
# now call that function on all the variable combinations
dfM <- sapply(vars, theCV, train=train, k=k, simplify = 'array', USE.NAMES = F)
df <- data.frame(formula = unlist(dfM[1,]),
MSE = unlist(dfM[2,]),
adj.R2 = unlist(dfM[3,]),
p.value = unlist(dfM[4,]), stringsAsFactors = F)
df
}
cv.lm('x7', train)
plus <- cv.lm(c('x7', 'x5', 'x3+x5'), train)
### FORWARD SUBSET SELECTION
FSS <- function(train, k) { #### the y variable MUST be called "y"
# master vector of variables
varsMaster <- names(train)[!grepl("y", names(train))]
# cross validation on single variables
df <- cv.lm(varsMaster, train, k)
# create a master df to store all levels
dfMaster <- df
# pick the best one
winner <- df[df$MSE==min(df$MSE), 1]
varsWinner <- gsub(" ", "", gsub("y ~ ", "", winner))
# subset out remaining vars
varsRemain <- varsMaster[!(varsMaster %in% unlist(strsplit(varsWinner, "+", fixed = T)))]
# paste remaining vars onto winners to create new combinations
newVars <- paste(varsWinner, varsRemain, sep="+")
#
# loop over all combinations, picking the best one each level
while(length(varsRemain) > 0) {
# run cross-validation with new variable combinations
df <- cv.lm(newVars, train, k)
# store new level stats in master df
dfMaster <- rbind(dfMaster, df)
# pick best one from new level
winner <- df[df$MSE==min(df$MSE), 1]
varsWinner <- gsub(" ", "", gsub("y ~ ", "", winner))
print(paste("varsWinner", varsWinner, df[df$MSE==min(df$MSE), 2])) #####
# subset out remaining vars
varsRemain <- varsMaster[!(varsMaster %in% unlist(strsplit(varsWinner, "+", fixed = T)))]
#print(paste("varsRemain", paste(varsRemain, collapse = ", "))) ###
# paste remaining vars onto winners to create new combinations
newVars <- paste(varsWinner, varsRemain, sep="+")
}
# output
print(paste("optimal model:", dfMaster[dfMaster$MSE == min(dfMaster$MSE), 1]))
list(dfMaster[dfMaster$MSE == min(dfMaster$MSE), 1], ## the optimal formula
dfMaster[dfMaster$MSE == min(dfMaster$MSE), ], ## the optimal formula plus stats for it
dfMaster) ## stats for all of the options tested
}
w <- FSS(train, 5)
w[[1]]
w[[2]]
# if you want to look choose by adj.R2 instead, just look at w[[3]] and pick the lowest adj.R2
### BACKWARD SUBSET SELECTION
BSS <- function(train, k) { #### the y variable MUST be called "y"
# master vector of variables
varsMaster <- names(train)[!grepl("y", names(train))]
# cross validation on all variables together
varsAll <- paste(varsMaster, collapse = "+")
df <- cv.lm(varsAll, train, k)
# create a master df to store all levels
dfMaster <- df
# pick the best one
winner <- varsAll
varsWinner <- gsub(" ", "", gsub("y ~ ", "", winner))
# subset out remaining vars
varsRemain <- unlist(strsplit(varsWinner, "+", fixed = T))
# paste remaining vars onto winners to create new combinations
newVars <- character()
for (i in 1:length(varsRemain)) {
newVars[i] <- paste(varsRemain[-i], collapse = "+")
}
#
# loop over all combinations, picking the best one each level
while(length(newVars) > 1) {
# run cross-validation with new variable combinations
df <- cv.lm(newVars, train, k)
# store new level stats in master df
dfMaster <- rbind(dfMaster, df)
# pick best one from new level
winner <- df[df$MSE==min(df$MSE), 1]
varsWinner <- gsub(" ", "", gsub("y ~ ", "", winner))
print(paste("varsWinner", varsWinner, df[df$MSE==min(df$MSE), 2])) ###
# make options for next level
varsRemain <- unlist(strsplit(varsWinner, "+", fixed = T))
newVars <- character()
for (i in 1:length(varsRemain)) {
newVars[i] <- paste(varsRemain[-i], collapse = "+")
}
}
# output
print(paste("optimal model:", dfMaster[dfMaster$MSE == min(dfMaster$MSE), 1]))
list(dfMaster[dfMaster$MSE == min(dfMaster$MSE), 1], ## the optimal formula
dfMaster[dfMaster$MSE == min(dfMaster$MSE), ], ## the optimal formula plus stats for it
dfMaster) ## stats for all of the options tested
}
bw <- BSS(train, 5)
bw[[1]]
bw[[2]]
# if you want to look choose by adj.R2 instead, just look at bw[[3]] and pick the lowest adj.R2
##### OTHER MODELS CROSS VALIDATION
## cross validated RANDOM FOREST
cv.rf <- function(train, k=5, returnDF=F) {
#make column for preds
train$preds <- factor(x=rep(levels(train$y)[1], nrow(train)),
levels = levels(train$y))
#randomize the indexes
nums <- sample(row.names(train), nrow(train))
#split the indexes into k groups
nv <- split(nums, cut(seq_along(nums), k, labels = FALSE))
#subset the training data into k folds
trainlist <- list()
for (i in 1:k) {
trainlist[[i]] <- train[nv[[i]], ]
}
#trainlist
#run on each fold
for (i in 1:k) {
ftrainlist <- trainlist[-i]
ftrain <- ftrainlist[[1]]
for (j in 2:length(ftrainlist)) {
ftrain <- rbind(ftrain, ftrainlist[[j]])
}
############# THE MODEL #######################
#mod <- lm(as.formula(paste(form,' - preds')), data = ftrain) ### the model
mod <- randomForest(y ~ .-preds, data=ftrain, importance=TRUE,proximity=FALSE) ### the model
###############################################
trainlist[[i]]$preds <- predict(mod, newdata = trainlist[[i]])
print(paste("finished fold", i))
}
#reassemble
cvdata <- ftrainlist[[1]]
for (j in 2:length(trainlist)) {
cvdata <- rbind(cvdata, trainlist[[j]])
}
# return stats
##raw accuracy
ra <- nrow(cvdata[cvdata$y == cvdata$preds,]) / nrow(cvdata)
print(paste("Raw Accuracy:", ra))
##balanced error rate
###http://spokenlanguageprocessing.blogspot.com/2011/12/evaluating-multi-class-classification.html
nk <- length(levels(train$y))
recall <- numeric(nk)
for (i in 1:nk) {
ck <- levels(train$y)[i]
recall[i] <- nrow(cvdata[cvdata$y==ck & cvdata$preds==ck,]) / nrow(cvdata[cvdata$y==ck,])
}
BER <- 1 - (sum(recall)/nk)
print(paste("Balanced Error Rate:", BER))
# return actual predictions
cvdata <- cvdata[order(as.numeric(row.names(cvdata))), ]
if(returnDF == T) {
return(cvdata[,c('y', 'preds')])
} else {
return()
}
}
predDF <- cv.rf(moddf, returnDF = T)
|
76353d72611753e502853ff241975c07be5c40ad
|
85df0b56b85eb23536cd7f95160b816884148238
|
/bnlearn.R
|
8731c0b2162e3edfe5ce0eba5f35b745b676ae7a
|
[] |
no_license
|
dgod1028/Bayesian-Network
|
5167c7c48256446af39f3e86d15446c1820bddc7
|
46da7afa9175af0321a82f06dbe470b2b432a142
|
refs/heads/master
| 2021-01-10T11:58:38.012575
| 2016-04-08T05:52:37
| 2016-04-08T05:52:37
| 54,442,688
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,819
|
r
|
bnlearn.R
|
## install.packages("bnlearn")
## 関連文献 http://arxiv.org/pdf/0908.3817.pdf
library(bnlearn)
library(Ecdat)
data(Fair)
head(Fair)
data <- Fair[,c(-2,-3,-9,-7)]
data[] <- lapply(data,as.factor)
colnames(data) <- c("性別","子供","宗教","教育","幸福")
bn.gs <- gs(data)
bn.gs
bn.hc <- hc(data,score ="aic")
bn.hc
par(mfrow = c(1,2))
plot(bn.gs, main = "Constraint-based algorithms")
plot(bn.hc, main = "Hill-Climbing",highlight = "幸福")
score <- score(bn.hc,data,type="aic")
score
fitted <- bn.fit(bn.hc,data,method="bayes") ### method = bayes or mle
(Coef <- coefficients(fitted))
Coef
Coef$幸福
#### Black list
banlist <- data.frame(from=c("性別","子供"),to=c("子供","性別"))
banlist
bn.hc2 <- hc(data,score="aic",blacklist = banlist)
bn.hc2
score2 <- score(bn.hc2,data,type="aic")
par(mfrow = c(1,1))
plot(bn.hc2, main = paste("Bayesian Network"," (Score: ",round(score2,digits=2),")",sep=""),,highlight = "幸福")
fitted2 <- bn.fit(bn.hc2,data,method="bayes") ### method = bayes or mle
(Coef2 <- coefficients(fitted2))
Coef2
Coef2$幸福
Coef2$教育
Coef2$宗教
Coef2$教育
#####
#### 作業中
#---絵に上書き
XMax <- max(axTicks(1))
YMax <- max(axTicks(2))
par(mfrow = c(1,1))
plot(bn.hc, main = "Hill-Climbing",highlight = "幸福")
#切片
text(XMax*0.05, YMax*0.7, round(Coef$性別, digits=2), adj=0)
text(XMax*0.85, YMax*0.6, round(Coef$SBP, digits=2), adj=0)
text(XMax*0.35, YMax*0.9, round(Coef$FBS, digits=2), adj=0)
text(XMax*0.35, YMax*0.05, round(Coef$BMI[1], digits=2), adj=0)
#回帰係数
text(XMax*0.3, YMax*0.3, round(Coef$BMI[2], digits=3), adj=0)
text(XMax*0.65, YMax*0.3, round(Coef$BMI[3], digits=3), adj=0)
text(XMax*0.45, YMax*0.5, round(Coef$BMI[4], digits=3), adj=0)
## [1] "性別" "子供" "宗教" "教育" "幸福"
|
0baf50d068a94b514b364113d45941efb6debeb9
|
bb789bd6b0649fae85c747710354122fd8ff7e25
|
/man/summary.pmlr.Rd
|
846a7cc219ab2298662057035554ca7df9777c4a
|
[] |
no_license
|
jshinb/pmlr
|
8efb9cf4e213f117db51ee4152747d9503408403
|
d000cc72192113356a84dd0ff501dae5fa98320f
|
refs/heads/master
| 2016-08-10T14:19:18.869369
| 2016-02-20T00:12:23
| 2016-02-20T00:12:23
| 49,685,089
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 1,376
|
rd
|
summary.pmlr.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/summary.pmlr.R
\name{summary.pmlr}
\alias{summary.pmlr}
\title{Summarizing Penalized Multinomial Logistic Model Fits}
\usage{
\method{summary}{pmlr}(object, ...)
}
\arguments{
\item{object}{an object of class \code{"pmlr"}, a result of a call to \code{\link{pmlr}}.}
\item{...}{further arguments passed to or from other methods}
}
\value{
\code{summary.pmlr} returns an object of class \code{"summary.pmlr"}, a list with components
\item{call}{the matched call of the \code{object}}
\item{method}{which method was used for hypothesis testing and computation of confidence intervals}
\item{coef}{an array containing the coefficient estimates, standard errors, and
test statistics and their p-values asssociated with the chosen method for the p parameters for the J categories}
\item{joint.test}{an array contatining the test statistics and p-values from constrained hypothesis tests under all betas = 0,
all betas are equal, and betas are proportional.}
\item{test.all0.vs.constraint}{Returned only if joint hypothesis testing was done:
An array containing likelihood ratio test statistics and p-values testing all \eqn{H_0}: betas=0 vs. other constraints (\eqn{H_C}),
which can be 'all betas are equal' or 'betas are proportion'.}
}
\description{
This function is for class \code{pmlr} object.
}
|
665e33bf4046de16fa0ff58b2c6f7ddc6002c1ab
|
fecb973f3ed39663ddeafeacac4b4e272c53f6fc
|
/R/mhealth.validate.R
|
53cce8389666033fbd64e3e84da8ec70fefe47f8
|
[
"MIT"
] |
permissive
|
qutang/mHealthR
|
79c173db45ea34b67566fbdd38fddc133b2e8621
|
4b3fef1fa24249a1c3c09d9721cd1148ae5c98a4
|
refs/heads/master
| 2018-12-09T04:44:39.771468
| 2018-09-11T20:09:01
| 2018-09-11T20:09:01
| 70,870,269
| 2
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 7,621
|
r
|
mhealth.validate.R
|
#' @name mhealth.validate
#' @title Validate filename or dataframe against mhealth specification
#' @param group_cols numeric or character vector specify columns to be validated as group variable. Only feasible for dataframe validation. Default is NULL, meaning there is no group columns in the dataframe.
#' @import stringr
#' @export
mhealth.validate = function(file_or_df, file_type, group_cols = NULL) {
if (is.character(file_or_df) & length(file_or_df) == 1) {
valid = .validate.filename(file_or_df, file_type)
} else if (is.data.frame(file_or_df)) {
# validate dataframe
valid = .validate.dataframe(file_or_df, file_type, group_cols)
} else{
valid = FALSE
message(
sprintf(
fmt = "\n
Input is not feasible for validation,
input class is: %s,
but should be filename string or dataframe",
class(file_or_df)
)
)
}
return(valid)
}
.validate.filename = function(filename, filetype) {
# validate number of sections
tokens = stringr::str_split(filename, "\\.", simplify = TRUE)
if (length(tokens) != 5 && length(tokens) != 6) {
valid = FALSE
message(
sprintf(
"\n
The number of sections separated by '.' is not correct: %s
There should be 'name', 'ID', 'timestamp', 'filetype', 'extension', 'opt-extension' sections.",
filename
)
)
return(valid)
}
# validate name section
valid = stringr::str_detect(tokens[1], pattern = mhealth$pattern$filename$NAME)
if (!valid) {
message(
sprintf(
"\n
Invalid name section: %s,
name section should only contain alphabets, numbers and '-'",
tokens[1]
)
)
}
# validate ID section
v_id = stringr::str_detect(tokens[2], pattern = mhealth$pattern$filename$ID)
valid = valid & v_id
if (!v_id) {
message(
sprintf(
"\n
Invalid id section: %s,
id section should only contain uppercase alphabets, numbers and '-'",
tokens[2]
)
)
}
# validate timestamp
ts = stringr::str_extract(filename, pattern = mhealth$pattern$filename$TIMESTAMP)
valid = valid &
.validate.timestamp(ts, format = mhealth$format$filename$TIMESTAMP)
# validate timezone
v_tz = stringr::str_detect(tokens[3], pattern = mhealth$pattern$filename$TIMEZONE)
valid = valid & v_tz
tz = stringr::str_extract(tokens[3], pattern = mhealth$pattern$filename$TIMEZONE)
if (!v_tz) {
message(sprintf(
"\n
Invalid time zone format: %s,
time zone should be like: [M/P]HHmm",
tz
))
}
# validate filetype
v_filetype = tokens[4] == filetype
valid = valid & v_filetype
if (!v_filetype) {
message(
sprintf(
"\n
Invalid file type section: %s,
file types should be %s",
tokens[4],
filetype
)
)
}
# validate extension
v_ext = tokens[5] == mhealth$pattern$filename$EXTENSION
valid = valid & v_ext
if (!v_ext) {
message(
sprintf(
"\n
Invalid extension: %s,
extension should be '%s'",
tokens[5],
mhealth$pattern$filename$EXTENSION
)
)
}
# validate optional extension
if (length(tokens) == 6) {
v_optext = tokens[6] == mhealth$pattern$filename$OPT_EXTENSION
valid = valid & v_optext
if (!v_optext) {
message(
sprintf(
"\n
Invalid optional extension: %s,
optional extension can only be '%s'",
tokens[6],
mhealth$pattern$filename$OPT_EXTENSION
)
)
}
}
# # validate overall pattern
# Use this simple method to valid as a whole pattern, but this way we can't get the insightful information where the pattern breaks
# valid = valid & stringr::str_detect(filename, mhealth$pattern$filename$FILENAME)
# if (!valid) {
# message(
# sprintf(
# "\n
# File name does not match mhealth convention: %s,
# As an example: %s",
# filename,
# mhealth$example$filename[filetype]
# )
# )
# }
return(valid)
}
.validate.dataframe = function(df, filetype, group_cols) {
required_cols = c(1)
cols = colnames(df)
ncols = ncol(df)
if(filetype == mhealth$filetype$annotation){
required_cols = 1:4
}else if(filetype == mhealth$filetype$feature){
required_cols = 1:3
}
# validate number of columns
if (filetype == mhealth$filetype$sensor ||
filetype == mhealth$filetype$event) {
valid = ncols >= 2
if (!valid) {
message(
sprintf(
"\n
The dataframe should have at least two columns for %s,
it only has: %d",
filetype,
ncols
)
)
return(valid)
}
}
if (filetype == mhealth$filetype$annotation ||
filetype == mhealth$filetype$feature) {
valid = ncols >= 4
if (!valid) {
message(
sprintf(
"\n
The dataframe should have at least fhour columns for %s,
it only has: %d",
filetype,
ncols
)
)
return(valid)
}
}
# validate column style
for (i in 1:ncols) {
valid = valid & .validate.columnstyle(cols, i)
}
# validate first column to be date
valid = .validate.columndate(df, 1)
# validate first column header
valid = valid &
.validate.columnheader(cols, 1, mhealth$column$TIMESTAMP, filetype)
# validate start and stop time column header for annotation and feature files
if (filetype == mhealth$filetype$feature ||
filetype == mhealth$filetype$annotation) {
valid = valid &
.validate.columnheader(cols, 2, mhealth$column$START_TIME, filetype)
valid = valid &
.validate.columnheader(cols, 3, mhealth$column$STOP_TIME, filetype)
# validate second and third column to be date for annotation and feature file
valid = valid & .validate.columndate(df, 2)
valid = valid & .validate.columndate(df, 3)
}
# validate the annotation name for annotation file
if(filetype == mhealth$filetype$annotation){
valid = valid & .validate.columnheader(cols, 4, mhealth$column$ANNOTATION_NAME, filetype)
valid = valid & .validate.columntype(df, 4, "character")
}
# convert group cols to numeric vector
if(is.character(group_cols)){
group_cols = sapply(group_cols, function(x){which(x == names(df))}, simplify = TRUE)
}else if(is.numeric(group_cols)){
}else{
message(sprintf(
"\n
group columns are not valid vector: %s
So it will be ignored",
class(group_cols)
))
group_cols = NULL
}
# validate group columns to be numeric or character
if(!is.null(group_cols)){
group_cols = setdiff(group_cols, required_cols)
result = sapply(group_cols, function(x){
if(x > ncol(df) || x < 1){
message(sprintf(
"\n
group column index %d does not exist",
x
))
return(FALSE)
}
if(!is.character(df[1,x]) && !is.numeric(df[1,x]) && !is.integer(df[1,x])){
message(sprintf(
"\n
group column %s is not in the correct format: %s",
names(df)[x], class(df[1, x])
))
return(FALSE)
}
return(TRUE)
})
valid = valid & all(result)
}
# validate numerical values for sensor type file
if (filetype == mhealth$filetype$sensor) {
validate_cols = 2:ncols
if(!is.null(group_cols)){
validate_cols = setdiff(2:ncols, group_cols)
}
for (i in validate_cols) {
valid = valid & .validate.columntype(df, i, "numeric")
}
}
return(valid)
}
|
af4a6ca8503482c042057826e8357c41876676a5
|
176deb3e42481c7db657cd945e2b53ad0dab66ca
|
/man/seg.Rd
|
ae4d2e4a51e74c2814c88ae688293f1c9a3336d7
|
[
"LicenseRef-scancode-public-domain-disclaimer",
"LicenseRef-scancode-warranty-disclaimer"
] |
permissive
|
katakagi/rloadest_test
|
a23d4ba635eadf00cebf5f9934217b3c5e16e0fd
|
74694ab65c7b62929961c45fdfa7eabf790869ef
|
refs/heads/master
| 2023-07-28T20:15:56.696712
| 2021-10-09T01:46:54
| 2021-10-09T01:46:54
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 567
|
rd
|
seg.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/seg.R
\name{seg}
\alias{seg}
\title{Load Model}
\usage{
seg(x, N)
}
\arguments{
\item{x}{the data to segment. Missing values are permitted and result
corresponing in missing values in output.}
\item{N}{the number of breaks in the segmented model.}
}
\value{
The data in \code{x} with attributes to build the segmented model.
}
\description{
Support function for building a segmented rating curve load model. Required
in the formula in \code{segLoadReg} to define the segmented model.
}
|
3e9b9ba36a41b00a8da687c2364d59bb41652cf2
|
4ba00ffc6623cdfc2d61f7b5e13918a889ee2472
|
/R/rbindfill-spdf.R
|
6448ed906db6aa61d41ed4adda8a728ff2e42398
|
[] |
no_license
|
fdetsch/StackExchange
|
acfe10041443c351a0336e38d62c72b0072ad90f
|
52dbd089da4fb9a59fcc3b40ebfcc2d129133b3e
|
refs/heads/main
| 2022-09-10T04:06:42.964264
| 2022-09-06T07:36:44
| 2022-09-06T07:36:44
| 99,312,967
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,092
|
r
|
rbindfill-spdf.R
|
### Answer to 'How to use rbind with SPDFs when the number of columns of arguments do not match?' -----
### (available online: https://gis.stackexchange.com/questions/267284/how-to-use-rbind-with-spdfs-when-the-number-of-columns-of-arguments-do-not-match/267355#267355)
library(raster)
France_map <- getData(name = "GADM", country = "FRA", level = 0, path = "D:/Data/GADM")
Germany_map <- getData(name = "GADM", country = "DEU", level = 1, path = "D:/Data/GADM")
Belgium_map <- getData(name = "GADM", country = "BEL", level = 2, path = "D:/Data/GADM")
## two-step approach
df = plyr::rbind.fill(France_map@data, Germany_map@data, Belgium_map@data)
spdf = SpatialPolygonsDataFrame(bind(as(France_map, "SpatialPolygons")
, as(Germany_map, "SpatialPolygons")
, as(Belgium_map, "SpatialPolygons"))
, data = df)
## all-in-one
out1 = bind(France_map, Germany_map, Belgium_map)
countries = list(France_map, Germany_map, Belgium_map)
out2 = do.call(bind, countries)
identical(out1, out2)
|
f0c57c6658d5effb36425ca452561d4c5bd8b404
|
d4edb03b1bf3b0a0fbc58208b6437fda4dbd3d6f
|
/inst/simulation_and_plotting_scripts/compare-methods_auRC-auPRC(discrete).R
|
749954556350da61ee528df407411960562edcb9
|
[
"MIT"
] |
permissive
|
insilico/npdr
|
245df976150f06ed32203354819711dfe60f211b
|
1aa134b6e82b700460e4b6392ef60ae2f80dfcfc
|
refs/heads/master
| 2023-06-23T20:24:58.581892
| 2023-06-10T00:53:23
| 2023-06-10T00:53:23
| 163,314,265
| 10
| 5
|
NOASSERTION
| 2021-09-21T18:15:34
| 2018-12-27T16:16:52
|
R
|
UTF-8
|
R
| false
| false
| 14,556
|
r
|
compare-methods_auRC-auPRC(discrete).R
|
# compute auRC and auPRC for NPDR, Relief, and Random Forest from 30 replicated data sets
# Discrete (SNP) Data
library(npdr)
library(CORElearn)
library(randomForest)
library(reshape2)
library(ggplot2)
library(PRROC)
show.plots = T # probalby want F for num.iter > 1 iterations
save.files = F # T for subsequent auPRC etc plots
run.pairs = T # compute all pairwise interaction stats and graph,
# probalby want F for num.iter > 1 iterations
num.iter <- 1 # just run one simulation
#num.iter <- 30 # 30 replicate simulations will take several minutes
if (save.files){
cat("Results files for ",num.iter, " replicate simulation(s) will be saved in ", getwd(),".", sep="")
}
# sim.type (options)
#
# "mainEffect": simple main effects
# "mainEffect_Erdos-Renyi": main effects with added correlation from Erdos-Renyi network
# "mainEffect_Scalefree": main effects with added correlation from Scale-free network
# "interactionErdos": interaction effects from Erdos-Renyi network
# "interactionScalefree": interaction effects from Scale-free network
# "mixed": main effects and interaction effects
# mix.type (options)
#
# "main-interactionErdos": main effects and interaction effects from Erdos-Renyi network
# "main-interactionScalefree": main effects and interaction effects from Scale-free network
# data.type (options)
#
# "continuous": random normal data N(0,1) (e.g., gene expression data)
# "discrete": random binomial data B(n=2,prob) (e.g., GWAS data)
num.samples <- 100
num.variables <- 100
main.bias <- 0.5
pct.mixed <- .5 # percent of effects that are main effects, must also use sim.type = "mixed"
pct.imbalance <- .5 # 0.25 => 75% case - 25% ctrl
mix.type <- "main-interactionErdos"
pct.signals <- 0.1
nbias <- round(pct.signals*num.variables)
sim.type <- "interactionErdos" # "mixed" for mixed main and interactions
data.type <- "discrete" # or "continuous" for standard normal data
auRC.npdr.multisurf <- numeric()
auRC.npdr.fixedk <- numeric()
auRC.relief <- numeric()
auRC.RF <- numeric()
auPRC.vec.npdr.multisurf <- numeric()
auPRC.vec.npdr.fixedk <- numeric()
auPRC.vec.relief <- numeric()
auPRC.vec.RF <- numeric()
set.seed(1989)
for(iter in 1:num.iter){
cat("Iteration: ",iter,"\n")
dataset <- createSimulation2(num.samples=num.samples,
num.variables=num.variables,
pct.imbalance=pct.imbalance,
pct.signals=pct.signals,
main.bias=main.bias,
interaction.bias=1, # 1/0 is max/min effect size
hi.cor=0.8,
lo.cor=0.1,
mix.type=mix.type,
label="class",
sim.type=sim.type,
pct.mixed=pct.mixed,
pct.train=0.5,
pct.holdout=0.5,
pct.validation=0,
plot.graph=FALSE,
verbose=TRUE,
use.Rcpp=FALSE,
prob.connected=0.3,
out.degree=(num.variables-2),
data.type=data.type)
dats <- rbind(dataset$train, dataset$holdout, dataset$validation)
dats <- dats[order(dats[,ncol(dats)]),]
if (run.pairs){ # only use this to explore one dataset
# pairwise interaction p-values or beta's if you want from inbixGAIN.R
output.type <- "Pvals"
intPairs.mat <- getInteractionEffects("class", dats,
regressionFamily = "binomial",
numCovariates = 0,
writeBetas = FALSE,
excludeMainEffects = FALSE,
interactOutput = output.type, # "Pvals", Betas", "stdBetas"
transformMethod = "",
numCores = 1,
verbose = F)
colnames(intPairs.mat) <- colnames(dats)[1:(ncol(dats)-1)]
rownames(intPairs.mat) <- colnames(dats)[1:(ncol(dats)-1)]
intPairs.Betas <- getInteractionEffects("class", dats,
regressionFamily = "binomial",
numCovariates = 0,
writeBetas = FALSE,
excludeMainEffects = FALSE,
interactOutput = "Betas", # "Pvals", Betas", "stdBetas"
transformMethod = "",
numCores = 1,
verbose = F)
colnames(intPairs.stdBetas) <- colnames(dats)[1:(ncol(dats)-1)]
rownames(intPairs.stdBetas) <- colnames(dats)[1:(ncol(dats)-1)]
want.to.adjust.p <- TRUE
if (output.type=="Pvals" & want.to.adjust.p){
# adjust p values
p.raw <- intPairs.mat[lower.tri(intPairs.mat, diag=FALSE)]
p.adj <- p.adjust(p.raw,method="fdr")
intPairs.mat[lower.tri(intPairs.mat, diag=FALSE)] <- p.adj
intPairs.mat <- intPairs.mat + t(intPairs.mat)
}
threshold <- .05
rm.nodes <- numeric()
A <- intPairs.mat
row <- 1
for (var in rownames(intPairs.mat)){
A[row,] <- 0 # make all A values 0 unless significant
if (output.type=="Pvals"){
signif.pairs.idx <- which(intPairs.mat[row,]>0 & intPairs.mat[row,]<threshold)
} else{ # matrix of betas
signif.pairs.idx <- which(intPairs.mat[row,]>threshold)
}
if (any(signif.pairs.idx)){
cat(var,": ", colnames(intPairs.mat)[signif.pairs.idx],"\n")
cat("pval.adj: ", intPairs.mat[row,signif.pairs.idx],"\n")
cat("beta: ", intPairs.Betas[row,signif.pairs.idx],"\n")
A[row,signif.pairs.idx] <- 1
} else{
# collect rows with no significant interactions for removal
rm.nodes <- c(rm.nodes,row)
}
row <- row + 1
}
A <- A[-rm.nodes,-rm.nodes] # remove nodes with no connections
# plot graph
g <- igraph::graph.adjacency(A)
# shape
V(g)$shape <- "circle"
V(g)$shape[grep("intvar",names(V(g)))] <- "rectangle"
# color
V(g)$color <- "gray"
V(g)$color[grep("intvar",names(V(g)))] <- "lightblue"
#igraph::V(g)$size <- scaleAB(degrees, 10, 20)
#png(paste(filePrefix, "_ba_network.png", sep = ""), width = 1024, height = 768)
plot(g, layout = igraph::layout.fruchterman.reingold, edge.arrow.mode = 0)
#vertex.size=(strwidth(V(g)$names) + strwidth("oo")) * 100,
#vertex.size2=strheight("I") * 100)
}
# npdr - multisurf
npdr.results1 <- npdr("class", dats, regression.type="binomial",
attr.diff.type="allele-sharing",
#nbd.method="relieff",
nbd.method="multisurf",
nbd.metric = "manhattan", msurf.sd.frac=.5, k=0,
neighbor.sampling="none", separate.hitmiss.nbds=F,
dopar.nn = T, dopar.reg=T, padj.method="bonferroni", verbose=T)
df1 <- data.frame(att=npdr.results1$att,
beta=npdr.results1$beta.Z.att,
pval=npdr.results1$pval.adj)
functional.vars <- dataset$signal.names
npdr.positives1 <- npdr.results1 %>% filter(pval.adj<.05) %>% pull(att)
npdr.positives1
npdr.results1[1:10,]
df1 <- na.omit(df1)
idx.func <- which(c(as.character(df1[,"att"]) %in% functional.vars))
func.betas1 <- df1[idx.func,"beta"]
neg.betas1 <- df1[-idx.func,"beta"]
pr.npdr1 <- PRROC::pr.curve(scores.class0 = func.betas1,
scores.class1 = neg.betas1,
curve = T)
if (show.plots){plot(pr.npdr1)}
npdr.detect.stats1 <- detectionStats(functional.vars, npdr.positives1)
# npdr - fixed k
npdr.results2 <- npdr("class", dats, regression.type="binomial",
attr.diff.type="allele-sharing",
nbd.method="relieff",
#nbd.method="multisurf",
nbd.metric = "manhattan", msurf.sd.frac=.5, k=0,
neighbor.sampling="none", separate.hitmiss.nbds=T,
dopar.nn = T, dopar.reg=T, padj.method="bonferroni", verbose=T)
npdr.positives2 <- npdr.results2 %>% filter(pval.adj<.05) %>% pull(att)
df2 <- data.frame(att=npdr.results2$att,
beta=npdr.results2$beta.Z.att,
pval=npdr.results2$pval.adj)
df2 <- na.omit(df2)
idx.func <- which(c(as.character(df2[,"att"]) %in% functional.vars))
func.betas2 <- df2[idx.func,"beta"]
neg.betas2 <- df2[-idx.func,"beta"]
pr.npdr2 <- PRROC::pr.curve(scores.class0 = func.betas2,
scores.class1 = neg.betas2,
curve = T)
if (show.plots){plot(pr.npdr2)}
npdr.detect.stats2 <- detectionStats(functional.vars, npdr.positives2)
#### Random Forest
ranfor.fit <- randomForest(as.factor(class) ~ ., data = dats)
rf.importance <- importance(ranfor.fit)
rf.sorted<-sort(rf.importance, decreasing=T, index.return=T)
rf.df <-data.frame(att=rownames(rf.importance)[rf.sorted$ix],rf.scores=rf.sorted$x)
rf.df[1:10,]
idx.func <- which(c(as.character(rf.df$att) %in% functional.vars))
func.scores.rf <- rf.df[idx.func,"rf.scores"]
neg.scores.rf <- rf.df[-idx.func,"rf.scores"]
pr.rf <- PRROC::pr.curve(scores.class0 = func.scores.rf,
scores.class1 = neg.scores.rf,
curve = T)
if (show.plots){plot(pr.rf)}
##### Regular Relief
relief <- CORElearn::attrEval(as.factor(class) ~ ., data = dats,
estimator = "ReliefFequalK",
costMatrix = NULL,
outputNumericSplits=FALSE,
kNearestEqual = floor(knnSURF(nrow(dats),.5)/2)) # fn from npdr
relief.order <- order(relief, decreasing = T)
relief.df <- data.frame(att=names(relief)[relief.order], rrelief=relief[relief.order])
idx.func <- which(c(as.character(relief.df$att) %in% functional.vars))
func.scores.relief <- relief.df[idx.func,"rrelief"]
neg.scores.relief <- relief.df[-idx.func,"rrelief"]
pr.relief <- PRROC::pr.curve(scores.class0 = func.scores.relief,
scores.class1 = neg.scores.relief,
curve = T)
if (show.plots){plot(pr.relief)}
pcts <- seq(0,1,.05)
rf.detected <- sapply(pcts,function(p){rfDetected(rf.df,functional.vars,p)})
relief.detected <- sapply(pcts,function(p){reliefDetected(relief.df,functional.vars,p)})
npdr.detected.multisurf <- sapply(pcts,function(p){npdrDetected(npdr.results1,functional.vars,p)})
npdr.detected.fixedk <- sapply(pcts,function(p){npdrDetected(npdr.results2,functional.vars,p)})
if (show.plots){
# plot recall curves (RC) for several methods
df <- data.frame(pcts=pcts, NPDR.MultiSURF=npdr.detected.multisurf,
NPDR.Fixed.k=npdr.detected.fixedk,
Relief=relief.detected,
RForest=rf.detected)
melt.df <- melt(data = df, measure.vars = c("NPDR.MultiSURF", "NPDR.Fixed.k", "Relief","RForest"))
gg <- ggplot(melt.df, aes(x=pcts, y=value, group=variable)) +
geom_line(aes(linetype=variable)) +
geom_point(aes(shape=variable, color=variable), size=2) +
scale_color_manual(values = c("#FC4E07", "brown", "#E7B800", "#228B22")) +
xlab("Percent Selected") +
ylab("Percent Correct") +
ggtitle("Power to Detect Functional Variables") +
theme(plot.title = element_text(hjust = 0.5)) + theme_bw()
print(gg)
# plot precision-recall curves (PRC) for several methods
method.vec <- c(rep('NPDR.MultiSURF',length=nrow(pr.npdr1$curve)),
rep('NPDR.Fixed.k',length=nrow(pr.npdr2$curve)),
rep('Relief',length=nrow(pr.relief$curve)),
rep('RForest',length=nrow(pr.rf$curve)))
df <- data.frame(recall=c(pr.npdr1$curve[,1],pr.npdr2$curve[,1],pr.relief$curve[,1],pr.rf$curve[,1]),
precision=c(pr.npdr1$curve[,2],pr.npdr2$curve[,2],pr.relief$curve[,2],pr.rf$curve[,2]),
method=method.vec)
gg <- ggplot(df, aes(x=recall, y=precision, group=method)) +
geom_line(aes(linetype=method)) +
geom_point(aes(shape=method, color=method), size=2) +
scale_color_manual(values= c("#FC4E07", "brown", "#E7B800", "#228B22")) +
xlab("Recall") +
ylab("Precision") +
ggtitle("Precision-Recall Curves: Comparison of Methods") +
theme(plot.title = element_text(hjust=0.5)) + theme_bw()
print(gg)
}
# auRC: area under the recall curve for several methods
auRC.RF[iter] <- sum(rf.detected)/length(rf.detected)
auRC.npdr.multisurf[iter] <- sum(npdr.detected.multisurf)/length(npdr.detected.multisurf)
auRC.npdr.fixedk[iter] <- sum(npdr.detected.fixedk)/length(npdr.detected.fixedk)
auRC.relief[iter] <- sum(relief.detected)/length(relief.detected)
# auPRC: area under the precision-recall curve for several methods
auPRC.vec.npdr.multisurf[iter] <- pr.npdr1$auc.integral
auPRC.vec.npdr.fixedk[iter] <- pr.npdr2$auc.integral
auPRC.vec.relief[iter] <- pr.relief$auc.integral
auPRC.vec.RF[iter] <- pr.rf$auc.integral
}
if (save.files){
# save results
df.save <- data.frame(cbind(iter=seq(1,30,by=1),
auRC.RForest=auRC.RF,
auRC.NPDR.MultiSURF=auRC.npdr.multisurf,
auRC.NPDR.Fixed.k=auRC.npdr.fixedk,
auRC.Relief=auRC.relief,
auPRC.RForest=auPRC.vec.RF,
auPRC.NPDR.MultiSURF=auPRC.vec.npdr.multisurf,
auPRC.NPDR.Fixed.k=auPRC.vec.npdr.fixedk,
auPRC.Relief=auPRC.vec.relief))
df.save <- apply(df.save,2,as.numeric)
#setwd("C:/Users/bdawk/Documents/KNN_project_output") will need to change to desired directory
file <- paste("auRC-auPRC_iterates_methods-comparison_",data.type,"-",sim.type,".csv",sep="")
write.csv(df.save,file,row.names=F)
}
|
711d25eca26fc012f7ba6995d5ffb63627896e21
|
3758b9c36518ec91a374660bc94745a2a615f675
|
/cmd/testsite/golden/load/test-library-4.0/fansi/tests/unitizer/_pre/funs.R
|
cff2b418d0422cd12a78c60f08533a4777a52597
|
[] |
no_license
|
metrumresearchgroup/pkgr
|
ac35bce7c4ae384daba0d6738fd52ef7b3ba16e8
|
27c581cb81b353769b88ea742e9649fbcc1b533d
|
refs/heads/develop
| 2023-06-08T20:07:04.191824
| 2023-06-06T14:00:06
| 2023-06-06T14:00:06
| 150,817,179
| 33
| 5
| null | 2023-08-30T14:06:50
| 2018-09-29T02:45:53
|
R
|
UTF-8
|
R
| false
| false
| 240
|
r
|
funs.R
|
## Helpers to extract the condition message only due to instability in
## C level error/warning in displaying the call or not
tce <- function(x) tryCatch(x, error=conditionMessage)
tcw <- function(x) tryCatch(x, warning=conditionMessage)
|
6f13a5979d7a7955926728bf01ebb12ff8f9e063
|
472eb2a35e6e58adddaf24dbd0b9aebf6aa61ad8
|
/02_nova_codificacao.R
|
02676227f26dad5ef6bf644f2ddb3cb50af46f48
|
[] |
no_license
|
neylsoncrepalde/midia-e-esfera-publica
|
99c04ce9c2074aedd12a4a56408beb908ed7f9d7
|
d08072cd592415b5e4f5f02c0d3ffacf426da86d
|
refs/heads/master
| 2021-09-15T07:03:49.802502
| 2018-05-28T07:23:23
| 2018-05-28T07:23:23
| 103,695,628
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 4,350
|
r
|
02_nova_codificacao.R
|
##############################
# Mídia e Esfera Pública
# Rousiley, Gabriella Hauber
# Script: Neylson Crepalde
##############################
setwd('~/Documentos/Rousiley')
list.files()
library(xlsx)
dados = read.xlsx("Nova_codificacao.xlsx",1)
View(dados)
nomes = names(dados)
nomes
nomes2 = gsub("X", "", nomes)
nomes2
names(dados) = nomes2
###########################################
library(reshape2)
library(dplyr)
dados$ocasiao = ifelse(dados$PH == 1, "Public Hearing", "Meeting")
names(dados)
emocoes = dados %>% select(1:6)
argumentos = dados %>% select(c(1, 14:37))
objetos = dados %>% select(c(1, 7:13))
names(argumentos)
emocoes_molten = melt(emocoes, id=1)
names(emocoes_molten)[2] = "emocao"
emocoes_molten = emocoes_molten %>% filter(value==1) %>% select(-value)
argumentos_molten = melt(argumentos, id=1)
names(argumentos_molten)[2] = "argumento"
argumentos_molten = argumentos_molten %>% filter(value==1) %>% select(-value)
objetos_molten = melt(objetos, id=1)
names(objetos_molten)[2] = "objeto"
objetos_molten = objetos_molten %>% filter(value==1) %>% select(-value)
molten = full_join(emocoes_molten, argumentos_molten, by='Argumentos')
emo_obj = full_join(emocoes_molten, objetos_molten, by="Argumentos")
obj_ocasiao = full_join(emo_obj, dados[c(1,40)], by="Argumentos")
molten_ocasiao = full_join(molten, dados[c(1,40)], by="Argumentos")
head(molten)
head(emo_obj)
head(molten_ocasiao)
# CRUZA EMOÇÕES E ARGUMENTOS
molten$argumento = factor(molten$argumento,
levels = c('1C','2C','3C','4C','5C','6C',
'7C','8C','9C','10C','11C','12C',
'1F','2F','3F','4F','5F','6F',
'7F','8F','9F','10F','11F','12F'))
table(molten$emocao, molten$argumento)
xtable::xtable(table(molten$emocao, molten$argumento))
# testa
fisher.test(table(molten$emocao, molten$argumento), simulate.p.value = T, B=5000)
chisq.test(table(molten$emocao, molten$argumento))
# PREPARANDO
indexc = grep("C", molten$argumento)
indexf = grep("F", molten$argumento)
molten$argumento = as.character(molten$argumento)
molten$posicionamento[indexc] = "Contra"
molten$posicionamento[indexf] = "A Favor"
contra = molten[indexc,]
afavor = molten[indexf,]
head(contra)
head(afavor)
# CRUZA EMOÇÕES E TIPO DE ARGUMENTO
table(molten$posicionamento, molten$emocao)
xtable::xtable(table(molten$posicionamento, molten$emocao))
chisq.test(table(molten$posicionamento, molten$emocao))
fisher.test(table(molten$posicionamento, molten$emocao))
# CRUZA EMOÇÕES E ARGUMENTOS
contra$argumento = factor(contra$argumento,
levels = c('1C','2C','3C','4C','5C','6C',
'7C','8C','9C','10C','11C','12C'))
#levels(contra$argumento) = c('1C','2C','3C','4C','5C','6C',
# '7C','8C','9C','10C','11C','12C')
table(contra$emocao, contra$argumento)
xtable::xtable(table(contra$emocao, contra$argumento))
chisq.test(table(contra$emocao, contra$argumento))
fisher.test(table(contra$emocao, contra$argumento), simulate.p.value = T, B=5000)
afavor$argumento = factor(afavor$argumento,
levels = c('1F','2F','3F','4F','5F','6F',
'7F','8F','9F','10F','11F','12F'))
xtable::xtable(table(afavor$emocao, afavor$argumento))
table(afavor$emocao, afavor$argumento)
fisher.test(table(afavor$emocao, afavor$argumento), simulate.p.value = T, B=5000)
# CRUZA EMOÇÕES E OBJETOS
table(emo_obj$emocao, emo_obj$objeto)
xtable::xtable(table(emo_obj$emocao, emo_obj$objeto))
fisher.test(table(emo_obj$emocao, emo_obj$objeto), simulate.p.value = T, B=5000)
# CRUZA EMOÇÕES E OCASIAO
table(molten_ocasiao$ocasiao, molten_ocasiao$emocao)
chisq.test(table(molten_ocasiao$ocasiao, molten_ocasiao$emocao))
# CRUZA POSICIONAMENTO E OCASIAO
table(molten_ocasiao$posicionamento, molten_ocasiao$ocasiao)
chisq.test(table(molten_ocasiao$posicionamento, molten_ocasiao$ocasiao))
# CRUZA OBJETOS E OCASIÃO
table(obj_ocasiao$ocasiao, obj_ocasiao$objeto)
xtable::xtable(table(obj_ocasiao$objeto, obj_ocasiao$ocasiao))
chisq.test(table(obj_ocasiao$ocasiao, obj_ocasiao$objeto))
fisher.test(table(obj_ocasiao$ocasiao, obj_ocasiao$objeto), simulate.p.value = T, B=5000)
######################################
|
60fa77a12c131eab107a8be989aab79fefefafd5
|
31698075d1580c6dc455adde3b30b636f0c87e70
|
/man/get_pdp_predictions.Rd
|
5c97d7bdfedc404424db58ff7d2d9c629bf75e4c
|
[] |
no_license
|
erblast/easyalluvial
|
11d10267e31ed44400f99c2e8d5df982ec90f468
|
df22644596db1eaa8e66f05e66b343de338d0d1f
|
refs/heads/master
| 2022-07-29T20:10:19.912023
| 2022-07-09T07:06:21
| 2022-07-09T07:06:21
| 149,339,634
| 101
| 10
| null | 2022-07-09T07:06:22
| 2018-09-18T19:13:25
|
R
|
UTF-8
|
R
| false
| true
| 3,757
|
rd
|
get_pdp_predictions.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/alluvial_model_response.R
\name{get_pdp_predictions}
\alias{get_pdp_predictions}
\title{get predictions compatible with the partial dependence plotting method}
\usage{
get_pdp_predictions(
df,
imp,
m,
degree = 4,
bins = 5,
.f_predict = predict,
parallel = FALSE
)
}
\arguments{
\item{df}{dataframe, training data}
\item{imp}{dataframe, with not more then two columns one of them numeric
containing importance measures and one character or factor column containing
corresponding variable names as found in training data.}
\item{m}{model object}
\item{degree}{integer, number of top important variables to select. For
plotting more than 4 will result in two many flows and the alluvial plot
will not be very readable, Default: 4}
\item{bins}{integer, number of bins for numeric variables, increasing this
number might result in too many flows, Default: 5}
\item{.f_predict}{corresponding model predict() function. Needs to accept `m`
as the first parameter and use the `newdata` parameter. Supply a wrapper for
predict functions with x-y syntax. For parallel processing the predict
method of object classes will not always get imported correctly to the worker
environment. We can pass the correct predict method via this parameter for
example randomForest:::predict.randomForest. Note that a lot of modeling
packages do not export the predict method explicitly and it can only be found
using :::.}
\item{parallel}{logical, turn on parallel processing. Default: FALSE}
}
\value{
vector, predictions
}
\description{
Alluvial plots are capable of displaying higher dimensional data
on a plane, thus lend themselves to plot the response of a statistical model
to changes in the input data across multiple dimensions. The practical limit
here is 4 dimensions while conventional partial dependence plots are limited
to 2 dimensions.
Briefly the 4 variables with the highest feature importance for a given
model are selected and 5 values spread over the variable range are selected
for each. Then a grid of all possible combinations is created. All
none-plotted variables are set to the values found in the first row of the
training data set. Using this artificial data space model predictions are
being generated. This process is then repeated for each row in the training
data set and the overall model response is averaged in the end. Each of the
possible combinations is plotted as a flow which is coloured by the bin
corresponding to the average model response generated by that particular
combination.
}
\details{
For more on partial dependency plots see
[https://christophm.github.io/interpretable-ml-book/pdp.html].
}
\section{Parallel Processing}{
We are using `furrr` and the `future` package to paralelize some of the
computational steps for calculating the predictions. It is up to the user
to register a compatible backend (see \link[future]{plan}).
}
\examples{
df = mtcars2[, ! names(mtcars2) \%in\% 'ids' ]
m = randomForest::randomForest( disp ~ ., df)
imp = m$importance
pred = get_pdp_predictions(df, imp
, m
, degree = 3
, bins = 5)
# parallel processing --------------------------
\dontrun{
future::plan("multisession")
# note that we have to pass the predict method via .f_predict otherwise
# it will not be available in the worker's environment.
pred = get_pdp_predictions(df, imp
, m
, degree = 3
, bins = 5,
, parallel = TRUE
, .f_predict = randomForest:::predict.randomForest)
}
}
|
4ba924dd9ed1e2bcd7e365f02cc8153f51670cf7
|
2f7eb4978331ab3585a57347f19a6a4323460191
|
/R/get_icc.R
|
962110b9cb99abc1f46465c797dd8e83c90308a0
|
[
"MIT"
] |
permissive
|
ernestguevarra/sampsizer
|
09ee112e7f9dc84cabab619167ceda4df0761a3e
|
801108c9c131b895465f3f328f04f642721953bd
|
refs/heads/master
| 2020-03-27T06:14:05.216847
| 2019-09-07T05:08:50
| 2019-09-07T05:08:50
| 146,090,847
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,039
|
r
|
get_icc.R
|
################################################################################
#
#'
#' Function to calculate the intracluster correlation coefficient of an
#' indicator collected from random cluster survey (RCS). This is a wrapper of
#' the \code{deff()} function in the \code{Hmisc} package.
#'
#' @param x variable to calculate ICC from
#' @param cluster variable identifying the clusters or groupings of the variable
#'
#' @return A vector with named elements n (total number of non-missing observations),
#' \code{clusters} (number of clusters after deleting missing data),
#' \code{rho} (intra-cluster correlation), and \code{deff} (design effect).
#'
#' @examples
#' x <- sample(1:2, size = 25, replace = TRUE)
#' cluster <- c(rep(1, 5), rep(2, 5), rep(3, 5), rep(4, 5), rep(5, 5))
#' get_icc(x = x, cluster = cluster)
#'
#' @export
#'
#
################################################################################
get_icc <- function(x, cluster) {
icc <- Hmisc::deff(y = x, cluster = cluster)
return(icc)
}
|
5d12096453c1c4d5ab8e2d156204435b6ba68949
|
5397b2f52030662f0e55f23f82e45faa165b8346
|
/R/j_get_meta.R
|
46c4c6392256e31d80e03c8d8bee3fd907a42d22
|
[
"MIT"
] |
permissive
|
data-science-made-easy/james-old
|
8569dcc8ce74c68bcbb81106127da4b903103fcd
|
201cc8e527123a62a00d27cd45d365c463fc1411
|
refs/heads/master
| 2023-01-12T21:16:23.231628
| 2020-11-19T13:46:17
| 2020-11-19T13:46:17
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 86
|
r
|
j_get_meta.R
|
#' @export j_get_meta
j_get_meta <- function(index) {
j_get(index, what = "meta")
}
|
342508343a5374389c1d68fa1d7e81b659cf13f1
|
4ff46765a3b93d40632c9c94d154b8b23e638fdd
|
/cba/machine-learning/codes/batch.R
|
37e7016c0f1aed3c613eae8a4f274ad976cb6556
|
[] |
no_license
|
harakonan/research-public
|
e5bedaa225aba44d1da624e9816434abe05529ac
|
5ea6a58e8603a08130095ad028c92ca791cbed3f
|
refs/heads/master
| 2023-06-28T04:27:35.057728
| 2023-06-08T11:47:55
| 2023-06-08T11:47:55
| 129,033,826
| 1
| 1
| null | null | null | null |
UTF-8
|
R
| false
| false
| 5,323
|
r
|
batch.R
|
# Batch file for data cleaning and analysis
# Do not forget to change set_env depending on the environment
# set_env <- "test"
set_env <- "main"
# Path to working directories
if (set_env == "test"){
source("~/Workspace/research-private/cba/machine-learning/codes/pathtowd/pathtowd_test.R")
} else if (set_env == "main"){
source("/mnt/d/userdata/k.hara/codes/pathtowd/pathtowd_main.R")
}
# track_r: Function that execute and track R script
source(paste0(pathtotools, "track_r.R"))
# Package loading
library(data.table)
library(zoo)
library(dplyr)
library(Hmisc)
library(ggplot2)
library(epiR)
library(pROC)
library(mnlogit)
library(splitstackshape)
library(doParallel)
library(mltools)
library(glmnet)
library(gam)
library(ranger)
library(xgboost)
library(nnet)
library(mda)
library(fastknn)
library(LiblineaR)
library(gridExtra)
# cba_data_cleaning.R
if (set_env == "test"){
claim_startmon <- 201404
claim_endmon <- 201603
hs_startday <- 20140401
hs_endday <- 20160331
agepoint <- 201603
track_r(pathtomain, "cba_data_cleaning_test.R", pathtolog, "cba_data_cleaning.log")
} else if (set_env == "main"){
claim_startmon <- 201604
claim_endmon <- 201803
hs_startday <- 20160401
hs_endday <- 20180331
agepoint <- 201803
track_r(pathtomain, "cba_data_cleaning.R", pathtolog, "cba_data_cleaning.log")
}
# cba_data_man_stat.R
if (set_env == "test"){
fyclaim <- 2015
} else if (set_env == "main"){
fyclaim <- 2017
}
track_r(pathtomain, "cba_data_man_stat.R", pathtolog, "cba_data_man_stat.log")
# cba_data_man_conv.R
if (set_env == "test"){
fyclaim <- 2015
} else if (set_env == "main"){
fyclaim <- 2017
}
track_r(pathtomain, "cba_data_man_conv.R", pathtolog, "cba_data_man_conv.log")
# Caution!: samples and labels will be refreshed if executed
# cba_sampling_refresh.R
sample_ratio <- NULL
target_disease <- "ht"
track_r(pathtomain, "cba_sampling_refresh.R", pathtolog, paste0("cba_sampling_refresh_", target_disease, ".log"))
sample_ratio <- NULL
target_disease <- "dm"
track_r(pathtomain, "cba_sampling_refresh.R", pathtolog, paste0("cba_sampling_refresh_", target_disease, ".log"))
sample_ratio <- NULL
target_disease <- "dl"
track_r(pathtomain, "cba_sampling_refresh.R", pathtolog, paste0("cba_sampling_refresh_", target_disease, ".log"))
# samples for testing cba_statlearn_analysis.R
sample_ratio <- 0.05
target_disease <- "ht"
track_r(pathtomain, "cba_sampling_refresh.R", pathtolog, paste0("cba_sampling_refresh_", target_disease, "_sample.log"))
sample_ratio <- 0.15
target_disease <- "dm"
track_r(pathtomain, "cba_sampling_refresh.R", pathtolog, paste0("cba_sampling_refresh_", target_disease, "_sample.log"))
sample_ratio <- 0.05
target_disease <- "dl"
track_r(pathtomain, "cba_sampling_refresh.R", pathtolog, paste0("cba_sampling_refresh_", target_disease, "_sample.log"))
# # cba_sampling_preserve.R
# sample_ratio <- NULL
# target_disease <- "ht"
# track_r(pathtomain, "cba_sampling_preserve.R", pathtolog, paste0("cba_sampling_preserve_", target_disease, ".log"))
# sample_ratio <- NULL
# target_disease <- "dm"
# track_r(pathtomain, "cba_sampling_preserve.R", pathtolog, paste0("cba_sampling_preserve_", target_disease, ".log"))
# sample_ratio <- NULL
# target_disease <- "dl"
# track_r(pathtomain, "cba_sampling_preserve.R", pathtolog, paste0("cba_sampling_preserve_", target_disease, ".log"))
# samples for testing cba_statlearn_analysis.R
# sample_ratio <- 0.05
# target_disease <- "ht"
# track_r(pathtomain, "cba_sampling_preserve.R", pathtolog, paste0("cba_sampling_preserve_", target_disease, "_sample.log"))
# sample_ratio <- 0.15
# target_disease <- "ht"
# track_r(pathtomain, "cba_sampling_preserve.R", pathtolog, paste0("cba_sampling_preserve_", target_disease, "_sample.log"))
# sample_ratio <- 0.05
# target_disease <- "dl"
# track_r(pathtomain, "cba_sampling_preserve.R", pathtolog, paste0("cba_sampling_preserve_", target_disease, "_sample.log"))
# cba_summary.R
if (set_env == "test"){
fyclaim <- 2015
} else if (set_env == "main"){
fyclaim <- 2017
}
track_r(pathtomain, "cba_summary.R", pathtolog, paste0("cba_summary", ".log"))
# cba_conventional_analysis.R
target_disease <- "ht"
track_r(pathtomain, "cba_conventional_analysis.R", pathtolog, paste0("cba_conventional_analysis_", target_disease, ".log"))
target_disease <- "dm"
track_r(pathtomain, "cba_conventional_analysis.R", pathtolog, paste0("cba_conventional_analysis_", target_disease, ".log"))
target_disease <- "dl"
track_r(pathtomain, "cba_conventional_analysis.R", pathtolog, paste0("cba_conventional_analysis_", target_disease, ".log"))
# cba_statlearn_analysis.R
# change full_data after test
target_disease <- "ht"
# full_data <- TRUE
full_data <- FALSE
track_r(pathtomain, "cba_statlearn_analysis.R", pathtolog, paste0("cba_statlearn_analysis_", target_disease, ".log"))
target_disease <- "dm"
# full_data <- TRUE
full_data <- FALSE
track_r(pathtomain, "cba_statlearn_analysis.R", pathtolog, paste0("cba_statlearn_analysis_", target_disease, ".log"))
target_disease <- "dl"
# full_data <- TRUE
full_data <- FALSE
track_r(pathtomain, "cba_statlearn_analysis.R", pathtolog, paste0("cba_statlearn_analysis_", target_disease, ".log"))
|
9fcc4793e8a7fdb8e83911aaf5c655161344402a
|
ac0a2b0c0dcbe8ea08d0baa42b66a38b2ffe6e37
|
/mean_substitution_rate_by_pango.R
|
8f2e6a84b30345c6d9836b849e4659e460632f4f
|
[
"CC0-1.0"
] |
permissive
|
cednotsed/ditto
|
3957e0fcb11ec505570bf7c6862a99c1916bb507
|
8a27ec1a06c2363ae36a6674457f3e7e2e1b9c9a
|
refs/heads/main
| 2023-04-17T18:43:09.591358
| 2022-05-08T03:50:28
| 2022-05-08T03:50:28
| 430,035,274
| 1
| 0
|
CC0-1.0
| 2022-05-08T03:50:30
| 2021-11-20T07:13:17
|
R
|
UTF-8
|
R
| false
| false
| 12,373
|
r
|
mean_substitution_rate_by_pango.R
|
rm(list = ls())
setwd("../Desktop/git_repos/ditto/")
require(tidyverse)
require(data.table)
require(ape)
require(ggplot2)
require(see)
require(foreach)
prefixes <- list.files("results/human_animal_subsets/V5/dating_out/")
ant_df <- read.csv("data/metadata/netherlands_humans_anthroponoses.txt", header = F)
prefixes <- prefixes[!grepl("deer|all_|USA|.png", prefixes)]
human_background <- "V5"
cluster_meta <- fread("results/cluster_annotation/deer_mink_parsed_clusters.csv") %>%
select(accession_id, cluster)
morsels <- foreach (prefix = prefixes) %do% {
country_name <- str_split(prefix, "\\.")[[1]][2]
time_tree <- read.nexus(str_glue("results/human_animal_subsets/{human_background}/dating_out/{prefix}/timetree.nexus"))
div_tree <- read.nexus(str_glue("results/human_animal_subsets/{human_background}/dating_out/{prefix}/divergence_tree.nexus"))
meta <- fread(str_glue("data/metadata/human_animal_subsets/{human_background}/{prefix}.csv")) %>%
select(-cluster) %>%
filter(accession_id %in% time_tree$tip.label) %>%
left_join(cluster_meta)
# Divide time tree by divergence tree
rate_tree <- div_tree
rate_tree$edge.length <- rate_tree$edge.length / time_tree$edge.length
# Drop anthroponotic tips
rate_tree <- drop.tip(rate_tree, ant_df$V1)
n <- length(rate_tree$tip.label)
rate_df <- tibble(accession_id = rate_tree$tip.label,
terminal_rate = rate_tree$edge.length[sapply(1:n, function(x,y) which(y == x),
y = rate_tree$edge[,2])])
rate_df %>%
left_join(meta) %>%
add_column(country = country_name) %>%
select(host, terminal_rate, cluster, country, pango_lineage)
}
plot_df <- bind_rows(morsels)
plot_df %>%
group_by(host, pango_lineage) %>%
summarise(n = n()) %>%
right_join(plot_df, by = c("host", "pango_lineage")) %>%
ggplot(aes(x = host, y = log(terminal_rate, base = 10), fill = host)) +
facet_grid(rows = vars(pango_lineage), scales = "free", space = "free") +
theme_bw() +
geom_point(color = "darkgrey", position = position_jitter(width = 0.08),
size = 0.5,
alpha = 0.8) +
geom_boxplot(position = position_nudge(x = -0.2, y = 0),
width = 0.2,
outlier.shape = NA,
alpha = 0.3) +
geom_violinhalf(position = position_nudge(x = 0.2, y = 0), alpha = 1) +
geom_text(aes(x = host, y = -1.5, label = paste0("n = ", n))) +
coord_flip() +
labs(y = "lg(subst. rate)", x = "Host") +
theme(legend.position = "none")
# ggsave(str_glue("results/human_animal_subsets/{human_background}/dating_out/mutations_rates_terminal.png"),
# width = 7,
# height = 5)
# Plot by cluster
plot_df %>%
group_by(cluster) %>%
summarise(n = n()) %>%
right_join(plot_df, by = c("cluster")) %>%
filter(host == "Neovison vison") %>%
ggplot(aes(x = cluster, y = log(terminal_rate, base = 10), fill = cluster)) +
facet_grid(rows = vars(country), scales = "free", space = "free") +
theme_bw() +
geom_point(color = "darkgrey", position = position_jitter(width = 0.08),
size = 0.5,
alpha = 0.8) +
geom_boxplot(position = position_nudge(x = -0.2, y = 0),
width = 0.2,
outlier.shape = NA,
alpha = 0.3) +
geom_violinhalf(position = position_nudge(x = 0.2, y = 0), alpha = 1) +
geom_text(aes(x = cluster, y = -1.5, label = paste0("n = ", n))) +
coord_flip() +
labs(y = "lg(subst. rate)", x = "Host") +
theme(legend.position = "none")
ggsave(str_glue("results/human_animal_subsets/{human_background}/dating_out/mutations_rates_terminal_by_cluster.png"),
width = 7,
height = 5)
#### ALL branches ###################
morsels2 <- foreach (prefix = prefixes) %do% {
country_name <- str_split(prefix, "\\.")[[1]][2]
time_tree <- read.nexus(str_glue("results/human_animal_subsets/{human_background}/dating_out/{prefix}/timetree.nexus"))
div_tree <- read.nexus(str_glue("results/human_animal_subsets/{human_background}/dating_out/{prefix}/divergence_tree.nexus"))
cluster_meta <- fread("results/cluster_annotation/deer_mink_parsed_clusters.csv") %>%
select(accession_id, cluster)
meta <- fread(str_glue("data/metadata/human_animal_subsets/{human_background}/{prefix}.csv")) %>%
filter(accession_id %in% time_tree$tip.label) %>%
left_join(cluster_meta)
human <- meta %>% filter(host == "Human")
animal <- meta %>% filter(host == "Neovison vison")
# Divide time tree by divergence tree
rate_tree <- div_tree
rate_tree$edge.length <- rate_tree$edge.length / time_tree$edge.length
# Drop anthroponotic tips
rate_tree <- drop.tip(rate_tree, ant_df$V1)
# Drop host tips
animal_tree <- drop.tip(rate_tree, human$accession_id)
human_tree <- drop.tip(rate_tree, animal$accession_id)
tibble(rate = animal_tree$edge.length, host = "Neovison vison") %>%
bind_rows(tibble(rate = human_tree$edge.length, host = "Human")) %>%
add_column(country = country_name) %>%
select(host, country, rate)
}
plot_df2 <- bind_rows(morsels2)
plot_df2 %>%
group_by(host, country) %>%
summarise(n = n()) %>%
right_join(plot_df2, by = c("host", "country")) %>%
ggplot(aes(x = host, y = log(rate, base = 10), fill = host)) +
facet_grid(rows = vars(country), scales = "free", space = "free") +
theme_bw() +
geom_point(color = "darkgrey", position = position_jitter(width = 0.08),
size = 0.5,
alpha = 0.8) +
geom_boxplot(position = position_nudge(x = -0.2, y = 0),
width = 0.2,
outlier.shape = NA,
alpha = 0.3) +
geom_violinhalf(position = position_nudge(x = 0.2, y = 0), alpha = 1) +
geom_text(aes(x = host, y = -1.5, label = paste0("n = ", n))) +
coord_flip() +
labs(y = "lg(subst. rate)", x = "Host") +
theme(legend.position = "none")
ggsave(str_glue("results/human_animal_subsets/{human_background}/dating_out/mutations_rates_internal.png"),
width = 7,
height = 5)
#### Specific mutations #####
morsels3 <- foreach (prefix = prefixes) %do% {
country_name <- str_split(prefix, "\\.")[[1]][2]
time_tree <- read.nexus(str_glue("results/human_animal_subsets/{human_background}/dating_out/{prefix}/timetree.nexus"))
div_tree <- read.nexus(str_glue("results/human_animal_subsets/{human_background}/dating_out/{prefix}/divergence_tree.nexus"))
cluster_meta <- fread("results/cluster_annotation/deer_mink_parsed_clusters.csv") %>%
select(accession_id, cluster)
meta <- fread(str_glue("data/metadata/human_animal_subsets/{human_background}/{prefix}.csv")) %>%
filter(accession_id %in% time_tree$tip.label) %>%
select(-cluster) %>%
left_join(cluster_meta)
# Divide time tree by divergence tree
rate_tree <- div_tree
rate_tree$edge.length <- rate_tree$edge.length / time_tree$edge.length
# Drop anthroponotic tips
rate_tree <- drop.tip(rate_tree, ant_df$V1)
n <- length(rate_tree$tip.label)
rate_df <- tibble(accession_id = rate_tree$tip.label,
terminal_rate = rate_tree$edge.length[sapply(1:n, function(x,y) which(y == x),
y = rate_tree$edge[,2])])
rate_df %>%
left_join(meta) %>%
add_column(country = country_name) %>%
select(host, terminal_rate, cluster, country)
}
plot_df3 <- bind_rows(morsels3) %>%
filter(host == "Neovison vison")
plot_df3 %>%
group_by(cluster) %>%
summarise(n = n()) %>%
right_join(plot_df3, by = "cluster") %>%
ggplot(aes(x = cluster, y = log(terminal_rate, base = 10), fill = cluster)) +
theme_bw() +
geom_point(color = "darkgrey", position = position_jitter(width = 0.08),
size = 0.5,
alpha = 0.8) +
geom_boxplot(position = position_nudge(x = -0.2, y = 0),
width = 0.2,
outlier.shape = NA,
alpha = 0.3) +
geom_violinhalf(position = position_nudge(x = 0.2, y = 0), alpha = 1) +
geom_text(aes(x = cluster, y = -1.5, label = paste0("n = ", n))) +
coord_flip() +
labs(y = "lg(subst. rate)", x = "Cluster") +
theme(legend.position = "none")
ggsave(str_glue("results/human_animal_subsets/{human_background}/dating_out/mutations_rates_terminal.png"),
width = 7,
height = 5)
##### Compute by mutation #####
mut_df <- fread("results/mink_homoplasy_alele_frequency_V5.csv") %>%
filter(mutation_annot %in% c("G37E", "Y486L", "N501T", "T229I", "L219V", "Y453F"))
# Load trees
prefix <- "all_mink.n1487.unambiguous.dedup"
time_tree <- read.nexus(str_glue("results/human_animal_subsets/{human_background}/dating_out/{prefix}/timetree.nexus"))
div_tree <- read.nexus(str_glue("results/human_animal_subsets/{human_background}/dating_out/{prefix}/divergence_tree.nexus"))
meta <- fread(str_glue("data/metadata/human_animal_subsets/{human_background}/{prefix}.csv")) %>%
filter(accession_id %in% time_tree$tip.label)
aln_prefix <- gsub(".unambiguous", ".audacity_only.v8_masked.unambiguous", prefix)
aln <- read.dna(str_glue("data/alignments/human_animal_subsets/{human_background}/{aln_prefix}.fasta"),
format = "fasta",
as.matrix = T)
morsels <- foreach (i = seq(nrow(mut_df))) %do% {
row <- mut_df[i, ]
# Get allele by position
aln_df <- tibble(accession_id = rownames(aln),
allele = toupper(as.vector(as.character(aln[, row$nucleotide_pos]))))
# Divide time tree by divergence tree
rate_tree <- div_tree
rate_tree$edge.length <- rate_tree$edge.length / time_tree$edge.length
# Drop anthroponotic tips
rate_tree <- drop.tip(rate_tree, ant_df$V1)
n <- length(rate_tree$tip.label)
with_df <- tibble(accession_id = rate_tree$tip.label,
terminal_rate = rate_tree$edge.length[sapply(1:n, function(x,y) which(y == x),
y = rate_tree$edge[,2])]) %>%
left_join(aln_df) %>%
filter(allele == row$var_nuc) %>%
add_column(mutation = row$mutation_annot,
mutation_present = "Present")
without_df <- tibble(accession_id = rate_tree$tip.label,
terminal_rate = rate_tree$edge.length[sapply(1:n, function(x,y) which(y == x),
y = rate_tree$edge[,2])]) %>%
left_join(aln_df) %>%
filter(allele != row$var_nuc) %>%
add_column(mutation = row$mutation_annot,
mutation_present = "Absent")
with_df %>%
bind_rows(without_df)
}
bind_rows(morsels) %>%
mutate(log_rate = log(terminal_rate, base = 10)) %>%
ggplot(aes(x = as.factor(mutation), y = log_rate, fill = mutation_present)) +
geom_violinhalf(position = position_dodge(width = 1), alpha = 1) +
geom_point(position = position_jitterdodge(dodge.width = 1, jitter.width = 0.07),
size = 0.5,
alpha = 0.3,
color = "black") +
geom_boxplot(position = position_dodge2(width = 1),
width = 0.2,
outlier.shape = NA,
alpha = 1)
ggpubr::ggarrange(plotlist = mut_plots)
pos_1 <- position_jitterdodge(
jitter.width = 0.25,
jitter.height = 0,
dodge.width = 0.9
)
bind_rows(morsels) %>%
ggplot(aes(x = mutation, y = terminal_rate, color = mutation_present)) +
# geom_jitter(alpha = 0.4, position = pos_1) +
stat_summary(position = position_dodge(width = 0.5),
fun.y = "mean",
geom = "point",
size = 3) +
stat_summary(position = position_dodge(width = 0.5),
fun.data = "mean_cl_normal",
geom = "errorbar",
width = 0.05,
lwd = 1,
fun.args = list(conf.int = 0.95)) +
labs(y = "Subst. rate", x = "Mutation", color = "")
ggsave(str_glue("results/human_animal_subsets/{human_background}/dating_out/mutations_rates_terminal_by_mutation.png"),
width = 7,
height = 5)
|
2907d4a548e7eed2a183281696420f5cebbb6704
|
50131539a2e92a690f43aea65b0c8ff10b90bbc0
|
/R/theme_coffee.R
|
7733fec4e244824f1792cbf47f85028c6cc88808
|
[
"MIT"
] |
permissive
|
RMHogervorst/coffeegeeks
|
904bc083fec2fc19e16151020795f19c33648698
|
23deb87e58029c6d345a5f1aa0e0ec07e96b86ca
|
refs/heads/master
| 2021-01-16T19:46:37.509578
| 2017-09-16T14:56:04
| 2017-09-16T14:56:04
| 100,188,183
| 1
| 1
| null | 2017-09-12T17:41:59
| 2017-08-13T16:01:48
|
CSS
|
UTF-8
|
R
| false
| false
| 2,344
|
r
|
theme_coffee.R
|
theme_coffee <- function(base_size=12, base_family="sans"){
ggplot2::theme(rect = element_rect(colour = "black", fill = "white"),
line = element_line(colour = "black"),
text = element_text(colour = "black"),
plot.title = element_text(face = "bold",
# 16 pt, bold, align left
size = rel(1.33), hjust = 0),
panel.background = element_rect(fill = NA, colour = NA),
panel.border = element_rect(fill = NA, colour = NA),
# 12 pt
axis.title = element_text(face = "italic"),
# 12 pt
axis.text = element_text(),
axis.line = element_line(colour = "black"),
axis.ticks = element_blank(),
panel.grid.major = element_line(colour = "#CCCCCC"),
panel.grid.minor = element_blank(),
legend.background = element_rect(colour = NA),
legend.key = element_rect(colour = NA),
legend.position = "right",
legend.direction = "vertical")
}
#' A set of coffee colors
#'
#' @rdname coffeetheme
#' @export
#' @inheritParams ggplot2::scale_colour_hue
coffee_pal <- function(){
function(n){
colors <- get_cols_data_frame()$hex_code[11:44]
colors[seq_len(n)]
}
}
#' @rdname coffeetheme
#' @export
scale_colour_coffee <- function(...){
discrete_scale("color", "coffeecolors", coffee_pal(), ...)
}
#' @rdname coffeetheme
#' @export
scale_color_coffee <- scale_colour_coffee
#' @rdname coffeetheme
#' @export
scale_fill_coffee <- function(...){
discrete_scale("fill", "coffeecolors", coffee_pal(), ...)
}
#' @rdname coffeetheme
#' @inheritParams ggplot2::scale_colour_hue
#' @export
scale_gradient_coffee <- function(...,
space = "Lab",
na.value = "grey50",
guide = "colourbar"){
scale_color_gradient(..., low = coffee_cols(cream),
high = coffee_cols(espresso),
space = space,
na.value = na.value,
guide = guide)
}
|
a3b3a40f870cbb34eded25f2b6edc0d190a243e0
|
770b14ae44e4991d444f0a0b1af124396bf2960f
|
/pkg/man/memisc-deprecated.Rd
|
c781641bbc0811b5941a701f27834785bd8ca0b0
|
[] |
no_license
|
melff/memisc
|
db5e2d685e44f3e2f2fa3d50e0986c1131a1448c
|
b5b03f75e6fe311911a552041ff5c573bb3515df
|
refs/heads/master
| 2023-07-24T19:44:10.092063
| 2023-07-07T23:09:11
| 2023-07-07T23:09:11
| 29,761,100
| 40
| 10
| null | 2022-08-19T19:19:13
| 2015-01-24T01:21:55
|
R
|
UTF-8
|
R
| false
| false
| 1,615
|
rd
|
memisc-deprecated.Rd
|
\name{memisc-deprecated}
\alias{memisc-deprecated}
\alias{fapply}
\alias{fapply.default}
\title{Deprecated Functions in Package \pkg{memisc}}
\description{
These functions are provided for compatibility with older versions of
\pkg{memisc} only, and may be defunct as soon as the next release.
}
\usage{
fapply(formula,data,\dots) # calls UseMethod("fapply",data)
\method{fapply}{default}(formula, data, subset=NULL,
names=NULL, addFreq=TRUE,\dots)
}
\arguments{
\item{formula}{a formula. The right hand side includes one or more
grouping variables separated by '+'. These may be factors, numeric,
or character vectors. The left hand side may be empty,
a numerical variable, a factor, or an expression.
See details below.}
\item{data}{an environment or data frame or an object coercable into a data frame.}
\item{subset}{an optional vector specifying a subset of observations
to be used.}
\item{names}{an optional character vector giving names to the
result(s) yielded by the expression on the left hand side of \code{formula}.
This argument may be redundant if the left hand side results in is a named vector.
(See the example below.)}
\item{addFreq}{a logical value. If TRUE and
\code{data} is a table or a data frame with a variable
named "Freq", a call to
\code{table}, \code{\link{Table}}, \code{\link{percent}}, or \code{\link{nvalid}}
is supplied by an additional argument \code{Freq}
and a call to \code{table} is translated into
a call to \code{Table}.
}
\item{\dots}{further arguments, passed to methods or ignored.}
}
|
a64bc677622f253be64e2a4663447c5f9c2fbb99
|
32ca51d8fb4de3e520b4aa49b545acefed74c3b2
|
/binder/install.R
|
303d08bd98f60deda45709203c155a7beaac6011
|
[] |
no_license
|
cannin/repo2docker-test
|
639211141beda5979a56536aaf653c1ea9eca445
|
1e7671acfbbfaa6855bdc19ec8a2f6de96306b2a
|
refs/heads/master
| 2020-05-18T09:33:28.996626
| 2019-08-05T17:54:24
| 2019-08-05T17:54:24
| 184,327,815
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 191
|
r
|
install.R
|
install.packages("devtools")
install.packages("rJava")
#source("https://gist.githubusercontent.com/cannin/6b8c68e7db19c4902459/raw/installPackages.R")
#installPackages("r-requirements.txt")
|
23590832b02d5139b5f6749c34fdb49d1afbc9d2
|
7f9d20b9e57be5dad4feeabf3e7f5fc6448c04e5
|
/WB_Data_Africa.R
|
bfe2b7a7c39f611a2792e5368938aaa08493c651
|
[] |
no_license
|
monmon1994/TB_COVID19_Africa
|
3de5e9d03cdea9c3c459be7271d9f7e814ac7220
|
2c4d01c2d211b9ed0cef4ac36c49580e785d23be
|
refs/heads/master
| 2022-07-20T22:12:59.814109
| 2020-05-13T08:48:22
| 2020-05-13T08:48:22
| 263,574,765
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 6,767
|
r
|
WB_Data_Africa.R
|
# COVID Data and World Bank Data for Africa
library(wbstats)
library(tidyverse)
library(tidycovid19)
library(scales)
library(ggrepel)
# World Bank Data
# Load the data from the stats
wb_africa <- wb(country = c("DZA", "AGO", "BEN", "BWA", "BFA", "BDI", "CV", "CMR", "CAF", "TCD", "COM", "COD",
"COG", "CIV", "DJI", "EGY", "ERI", "SWZ", "ETH", "GAB", "GMB", "GHA", "GIN", "GNB",
"KEN", "LSO", "LBR", "LBY", "MDG", "MWI", "MLI", "MRT", "MUS", "MAR", "MOZ", "NAM",
"NER", "NGA", "RWA", "STP", "SEN", "SYC", "SLE", "SOM", "ZAF", "SSD", "SDN", "TZA",
"TGO", "TUN", "UGA", "ZMB", "ZWE"),
indicator = c("SH.MED.PHYS.ZS", "SH.XPD.CHEX.GD.ZS", "SH.XPD.OOPC.CH.ZS",
"SH.STA.DIAB.ZS", "SH.PRV.SMOK", "SH.DYN.NCOM.ZS", "SH.STA.BASS.ZS",
"SP.POP.65UP.TO.ZS", "NY.GDP.MKTP.CD", "EN.POP.DNST", "SH.MED.BEDS.ZS",
"SH.STA.HYGN.ZS", "NY.GDP.PCAP.CD"),
startdate = 2015,
enddate = 2018, POSIXct = T, return_wide = T, removeNA = T)
## COVID DATA AFRICA
df <- download_jhu_csse_covid19_data(silent = T, cached = T)
df <- df[['country']]
df_africa <- df %>%
filter(iso3c %in% c("DZA", "AGO", "BEN", "BWA", "BFA", "BDI", "CV", "CMR", "CAF", "TCD", "COM", "COD",
"COG", "CIV", "DJI", "EGY", "ERI", "SWZ", "ETH", "GAB", "GMB", "GHA", "GIN", "GNB",
"KEN", "LSO", "LBR", "LBY", "MDG", "MWI", "MLI", "MRT", "MUS", "MAR", "MOZ", "NAM",
"NER", "NGA", "RWA", "STP", "SEN", "SYC", "SLE", "SOM", "ZAF", "SSD", "SDN", "TZA",
"TGO", "TUN", "UGA", "ZMB", "ZWE")) %>%
group_by(country) %>%
mutate(
total_confirmed = max(confirmed),
total_deaths = max(deaths)
) %>%
select(country, total_confirmed, total_deaths, iso3c) %>%
distinct() %>%
ungroup() %>%
arrange(-total_deaths)
merged <- merge(wb_africa, df_africa, by = "iso3c") # merge the df's
save(merged, file = "data/merged.RData")
########### plots
pal <- wesanderson::wes_palette("BottleRocket2", n = 51, type = "continuous")
# AGE 65 above
merged %>%
filter(date == 2018) %>%
ggplot() +
geom_point(aes(x = SP.POP.65UP.TO.ZS, y = total_confirmed, colour = country.x, size = 3),
show.legend = FALSE) +
scale_fill_gradientn(colours = pal) +
scale_y_log10() +
geom_label_repel(aes(x = SP.POP.65UP.TO.ZS, y = total_confirmed, label = country.x)) +
guides(fill=FALSE) +
xlab("Population ages 65 and above (% of total)") +
ylab("COVID-19 Confirmed Cases") +
theme_classic()
# SMOKING SH.PRV.SMOK
ggplot(merged) +
geom_point(aes(x = SH.PRV.SMOK, y = total_confirmed, colour = country.x, size = 3),
show.legend = FALSE) +
scale_fill_gradientn(colours = pal) +
scale_y_log10() +
geom_label_repel(aes(x = SH.PRV.SMOK, y = total_confirmed, label = country.x)) +
guides(fill=FALSE) +
xlab("Smoking prevalence, total, ages 15+") +
ylab("COVID-19 Confirmed Cases") +
theme_classic()
# Physicians per 1,000 SH.MED.PHYS.ZS
ggplot(merged) +
geom_point(aes(x = SH.MED.PHYS.ZS, y = total_confirmed, colour = country.x, size = 3),
show.legend = FALSE) +
scale_fill_gradientn(colours = pal) +
scale_y_log10() +
geom_label_repel(aes(x = SH.MED.PHYS.ZS, y = total_confirmed, label = country.x)) +
guides(fill=FALSE) +
xlab("Physicians (1,000 per people)") +
ylab("COVID-19 Confirmed Cases") +
theme_classic()
# Non-communicable diseases SH.DTH.NCOM.ZS
ggplot(merged) +
geom_point(aes(x = SH.DYN.NCOM.ZS, y = total_confirmed, colour = country.x, size = 3),
show.legend = FALSE) +
scale_fill_gradientn(colours = pal) +
scale_y_log10() +
geom_label_repel(aes(x = SH.DYN.NCOM.ZS, y = total_confirmed, label = country.x)) +
guides(fill=FALSE) +
xlab("Mortality from CVD, cancer, diabetes or
CRD between exact ages 30 and 70 (%)") +
ylab("COVID-19 Confirmed Cases") +
theme_classic()
# SH.STA.DIAB.ZS no data
# SH.STA.HYGN.ZS basic hygiene services
merged %>%
filter(date == 2017) %>%
ggplot() +
geom_point(aes(x = SH.STA.HYGN.ZS, y = total_confirmed, colour = country.x, size = 3),
show.legend = FALSE) +
scale_fill_gradientn(colours = pal) +
scale_y_log10() +
guides(fill=FALSE, size = F) +
gghighlight::gghighlight(use_direct_label = T) +
xlab("People with basic handwashing facilities including soap and water (% of population)") +
ylab("COVID-19 Confirmed Cases") +
theme_classic()
# NY.GDP.MKTP.CD
merged %>%
filter(date == 2018) %>%
ggplot() +
geom_point(aes(x = NY.GDP.MKTP.CD, y = total_confirmed, colour = country.x, size = 3),
show.legend = FALSE) +
scale_fill_gradientn(colours = pal) +
scale_y_log10() +
scale_x_log10() +
guides(fill=FALSE, size = F) +
gghighlight::gghighlight(use_direct_label = T) +
xlab("GDP (current USD)") +
ylab("COVID-19 Confirmed Cases") +
theme_classic()
# NY.GDP.PCAP.CD GDP per capita
merged %>%
filter(date == 2018) %>%
ggplot() +
geom_point(aes(x = NY.GDP.PCAP.CD, y = total_confirmed, colour = country.x, size = 3),
show.legend = FALSE) +
scale_fill_gradientn(colours = pal) +
scale_y_log10() +
scale_x_log10() +
guides(fill=FALSE, size = F) +
gghighlight::gghighlight(use_direct_label = T) +
xlab("GDP (current USD)") +
ylab("COVID-19 Confirmed Cases") +
theme_classic()
# Health expenditure "SH.XPD.CHEX.GD.ZS"
merged %>%
filter(date == 2017) %>%
ggplot() +
geom_point(aes(x = SH.XPD.CHEX.GD.ZS, y = total_confirmed, colour = country.x, size = 3),
show.legend = FALSE) +
scale_fill_gradientn(colours = pal) +
scale_y_log10() +
guides(fill=FALSE, size = F) +
gghighlight::gghighlight(use_direct_label = T) +
xlab("Current health expenditure (% of GDP)") +
ylab("COVID-19 Confirmed Cases") +
theme_classic()
# SH.XPD.OOPC.CH.ZS "Out-of-pocket expenditure (% of current health expenditure)"
merged %>%
filter(date == 2017) %>%
ggplot() +
geom_point(aes(x = SH.XPD.OOPC.CH.ZS, y = total_confirmed, colour = country.x, size = 3),
show.legend = FALSE) +
scale_fill_gradientn(colours = pal) +
scale_y_log10() +
guides(fill=FALSE, size = F) +
gghighlight::gghighlight(use_direct_label = T) +
xlab("Out-of-pocket expenditure (% of current health expenditure)") +
ylab("COVID-19 Confirmed Cases") +
theme_classic()
|
5bb7c25110290707bcbd78341103071b8449100a
|
0e6f03600e49c8a9f6bfe1c4e0ed2fb423d9690d
|
/Real_study/r/1_lacpd_wadi.R
|
aad41b3144521915e53bcfa547bc74ae9e2ad778
|
[] |
no_license
|
mmontesinosanmartin/LACPD_Article
|
bd1688dea09d993dcd859b01741a9c3449d28ee5
|
b8acd76eae1ea704c21cf8c9306dc34403e4f071
|
refs/heads/master
| 2023-02-16T04:15:10.784437
| 2021-01-12T17:48:11
| 2021-01-12T17:48:11
| 259,288,942
| 0
| 1
| null | null | null | null |
UTF-8
|
R
| false
| false
| 2,680
|
r
|
1_lacpd_wadi.R
|
###############################################################################
# R CODE: Locally adaptive change point detection
###############################################################################
# Moradi, M., Montesino-SanMartin, M., Ugarte, M.D., Militino, A.F.
# Public University of Navarre
# License: Availability of material under
# [CC-BY-SA](https://creativecommons.org/licenses/by-sa/2.0/).
###############################################################################
# PACKAGE
###############################################################################
# devtools package
# library(devtools)
# install from github
# install_github("mmontesinosanmartin/LACPD_Article/LACPD")
# load
library(LACPD)
###############################################################################
# DATA PATHS
###############################################################################
# inputs
root.git <- "https://raw.githubusercontent.com/mmontesinosanmartin/changepoint_article/"
root.dir <- "master/Real_study/data"
files <- c("field1_7.RData",
"field2_21.RData",
"field3_29.RData")
n.files <- length(files)
# output directory (select your own)
out.dir <- "./"
###############################################################################
# ANALYSIS
###############################################################################
# Paramters
# windows
this.k <- "02:10"
# iterations
this.m <- 100
# adjusting method
this.adj <- "BY"
# Check running times
t.strt <- Sys.time()
# Run: for each file
for(i in 1:n.files){
# load the dataset
f.i <- file.path(root.git, root.dir, paste0(files[i]))
load(url(f.i))
# result
resl <- list()
# for each pixel
for (j in 1:nrow(this.data)) {
# the time-series for this pixel
x <- this.data[j,]
# print message (every 10% completed)
p <- round(j/nrow(this.data)*100)
if (p %% 10 == 0 & p > 0) print(paste0(p, "%"))
# run the analysis
resl[[j]] <- lacpd_mk(x = x,
k = eval(parse(text=this.k)),
m = this.m,
adjust = TRUE,
method = this.adj,
history = TRUE)
}
# save output
suffix <- paste(gsub(":","",this.k), this.m, this.adj, sep = "_")
out.file <- paste0(gsub(".RData","",basename(f.i)),"_",suffix, ".RData")
out.path <- file.path(out.dir,out.file)
save(sample.roi, # study region
sample.val, # NDVI RasterBrick
this.data, # NDVI data matrix
resl, # All results
file = out.path)
}
# Total running time
Sys.time() - t.strt
# Time difference of 1.734562 hours
|
0b5fcd5e86a49b416e21482e56be05a22beecd1e
|
2970a3fe4634b1f8fb24243730b8becce4b5da42
|
/week_1/5_friday/Phylacine/commands.r
|
cf867f8bf0490e9bef76d4e80bbfa95f400f5335
|
[
"MIT"
] |
permissive
|
chase-lab/biodiv-patterns-course
|
21b38d5b99e3a90cc18e020c2fb74db9a5a8966b
|
8212e9fabb2db9187c510e6e15ce8e94e83f960e
|
refs/heads/master
| 2022-04-05T06:07:53.110753
| 2020-02-21T16:11:04
| 2020-02-21T16:11:04
| 235,298,888
| 2
| 1
| null | null | null | null |
UTF-8
|
R
| false
| false
| 467
|
r
|
commands.r
|
library(rgdal)
library(sp)
library(raster)
# ------------------------------------------------------------------------------
sh <- readOGR(dsn= "Shapefile", layer = "land_boundary")
rs <- raster("Environment/Elevation.tif")
plot(rs)
plot(sh, add=T)
# projections
BEHRMANN <- CRS("+proj=cea +lon_0=0 +lat_ts=30 +x_0=0 +y_0=0 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0")
WGS84 <- CRS("+proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0")
|
17b8ef6f3dcae7519cebb932a446dd011848db4a
|
0accbe1994d882fdeb49cad9398519cfb4e40a15
|
/r_scripts/170220_pol2_profiles_meta_meta.R
|
1a3219ce6cc7ffb18f3588d907aa5547ab544545
|
[
"MIT"
] |
permissive
|
linlabbcm/RASMC_Phenotypic_Switching
|
026da5f96dc0aef3705eb0218fd7d2d3990f025e
|
0636cf9154bae385c998e1a44c360f726e09ea56
|
refs/heads/master
| 2020-03-22T01:05:36.876041
| 2018-06-30T23:43:42
| 2018-06-30T23:43:42
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 8,457
|
r
|
170220_pol2_profiles_meta_meta.R
|
pol2_0_TSS = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TSS_ALL_-3000_+0\\RN6_TSS_ALL_-3000_+0_RASMC_POL2_UNSTIM_NEW.gff',header=TRUE)
pol2_0_TXN = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TXN_ALL_-0_+0\\RN6_TXN_ALL_-0_+0_RASMC_POL2_UNSTIM_NEW.gff',header=TRUE)
pol2_0_TTR = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TTR_ALL_-0_+3000\\RN6_TTR_ALL_-0_+3000_RASMC_POL2_UNSTIM_NEW.gff',header=TRUE)
pol2_2_TSS = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TSS_ALL_-3000_+0\\RN6_TSS_ALL_-3000_+0_RASMC_POL2_PDGF_2H_NEW.gff',header=TRUE)
pol2_2_TXN = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TXN_ALL_-0_+0\\RN6_TXN_ALL_-0_+0_RASMC_POL2_PDGF_2H_NEW.gff',header=TRUE)
pol2_2_TTR = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TTR_ALL_-0_+3000\\RN6_TTR_ALL_-0_+3000_RASMC_POL2_PDGF_2H_NEW.gff',header=TRUE)
pol2_24_TSS = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TSS_ALL_-3000_+0\\RN6_TSS_ALL_-3000_+0_RASMC_POL2_PDGF_24H_NEW.gff',header=TRUE)
pol2_24_TXN = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TXN_ALL_-0_+0\\RN6_TXN_ALL_-0_+0_RASMC_POL2_PDGF_24H_NEW.gff',header=TRUE)
pol2_24_TTR = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TTR_ALL_-0_+3000\\RN6_TTR_ALL_-0_+3000_RASMC_POL2_PDGF_24H_NEW.gff',header=TRUE)
pol2_2J_TSS = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TSS_ALL_-3000_+0\\RN6_TSS_ALL_-3000_+0_RASMC_POL2_PDGF_2H_JQ1_NEW.gff',header=TRUE)
pol2_2J_TXN = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TXN_ALL_-0_+0\\RN6_TXN_ALL_-0_+0_RASMC_POL2_PDGF_2H_JQ1_NEW.gff',header=TRUE)
pol2_2J_TTR = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TTR_ALL_-0_+3000\\RN6_TTR_ALL_-0_+3000_RASMC_POL2_PDGF_2H_JQ1_NEW.gff',header=TRUE)
pol2_24J_TSS = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TSS_ALL_-3000_+0\\RN6_TSS_ALL_-3000_+0_RASMC_POL2_PDGF_24H_JQ1_NEW.gff',header=TRUE)
pol2_24J_TXN = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TXN_ALL_-0_+0\\RN6_TXN_ALL_-0_+0_RASMC_POL2_PDGF_24H_JQ1_NEW.gff',header=TRUE)
pol2_24J_TTR = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\mappedFolder\\RN6_TTR_ALL_-0_+3000\\RN6_TTR_ALL_-0_+3000_RASMC_POL2_PDGF_24H_JQ1_NEW.gff',header=TRUE)
active_genes = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\activeListTable.txt', header = FALSE)
orderTable = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\170206_waterfall_genes_log2_24_v_0_ordered_BRD4_H3k_RNA_0_2_24_and_log2.txt',header = TRUE)
up_24=which(orderTable[,12]>=1)
down_24 = which(orderTable[,12]<=-1)
orderTable2 = read.delim('C:\\Users\\rhirsch\\Documents\\rasmc_docs\\170206_waterfall_genes_log2_2_v_0_ordered_BRD4_H3k_RNA_0_2_24_and_log2.txt',header = TRUE)
up_2=which(orderTable2[,10]>=1)
down_2 = which(orderTable2[,10]<=-1)
pol2_0 = cbind(pol2_0_TSS[,3:62],pol2_0_TXN[,3:202],pol2_0_TTR[,3:62])
rownames(pol2_0) = pol2_0_TTR[,1]
pol2_2 = cbind(pol2_2_TSS[,3:62], pol2_2_TXN[,3:202],pol2_2_TTR[,3:62])
rownames(pol2_2) = rownames(pol2_0)
pol2_24 = cbind(pol2_24_TSS[,3:62],pol2_24_TXN[,3:202],pol2_24_TTR[,3:62])
rownames(pol2_24) = pol2_0_TTR[,1]
pol2_2J = cbind(pol2_2J_TSS[,3:62], pol2_2J_TXN[,3:202],pol2_2J_TTR[,3:62])
rownames(pol2_2J) = rownames(pol2_0)
pol2_24J = cbind(pol2_24J_TSS[,3:62], pol2_24J_TXN[,3:202],pol2_24J_TTR[,3:62])
rownames(pol2_24J) = rownames(pol2_0)
###########################################
############Active Genes Metas#############
###########################################
active0 = c()
for(i in 1:length(active_genes[,1])){
row = which(as.character(rownames(pol2_0))==as.character(active_genes[i,1]))
active0=c(active0,row)
}
pol2_0_meta = apply(pol2_0[active0,],2,mean,na.rm=TRUE)
pol2_2_meta = apply(pol2_2[active0,],2,mean,na.rm=TRUE)
pol2_24_meta = apply(pol2_24[active0,],2,mean,na.rm=TRUE)
pol2_2J_meta = apply(pol2_2J[active0,],2,mean,na.rm=TRUE)
pol2_24J_meta = apply(pol2_24J[active0,],2,mean,na.rm=TRUE)
pdf(file='C:\\Users\\rhirsch\\Documents\\rasmc_figures_1-3\\170222_rasmc_active_pol2_0v2_meta.pdf',width = 10,height = 8)
plot(1:320, pol2_0_meta,type='l',col='red',ylim =c(0,2.5),xaxt='n',xlab='',ylab='rpm/bp',main='active genes 0H (red) vs. 2H (black)')
lines(1:320, pol2_2_meta,type='l',col='black')
axis(1,c(0,60,260,320),c('-3kb','TSS','END','+3kb'))
dev.off()
pdf(file='C:\\Users\\rhirsch\\Documents\\rasmc_figures_1-3\\170222_rasmc_active_pol2_0v24_meta.pdf',width = 10,height = 8)
plot(1:320, pol2_0_meta,type='l',col='red',ylim =c(0,2.5),xaxt='n',xlab='',ylab='rpm/bp',main='active genes 0H (red) vs. 24H (black)')
lines(1:320, pol2_24_meta,type='l',col='black')
axis(1,c(0,60,260,320),c('-3kb','TSS','END','+3kb'))
dev.off()
pdf(file='C:\\Users\\rhirsch\\Documents\\rasmc_figures_1-3\\170222_rasmc_active_pol2_2v2J_meta.pdf',width = 10,height = 8)
plot(1:320, pol2_2_meta,type='l',col='red',ylim =c(0,2.5),xaxt='n',xlab='',ylab='rpm/bp',main='active genes 2H (red) vs. 2H+JQ1 (black)')
lines(1:320, pol2_2J_meta,type='l',col='black')
axis(1,c(0,60,260,320),c('-3kb','TSS','END','+3kb'))
dev.off()
pdf(file='C:\\Users\\rhirsch\\Documents\\rasmc_figures_1-3\\170222_rasmc_active_pol2_24v24J_meta.pdf',width = 10,height = 8)
plot(1:320, pol2_24_meta,type='l',col='red',ylim =c(0,5),xaxt='n',xlab='',ylab='rpm/bp',main='active genes 24H (red) vs. 24H+JQ1 (black)')
lines(1:320, pol2_24J_meta,type='l',col='black')
axis(1,c(0,60,260,320),c('-3kb','TSS','END','+3kb'))
dev.off()
###########################################
############Gained/Lost Genes Metas 24H####
###########################################
lost_genes=c()
for(i in 1:1000){
row = which(as.character(active_genes[,2]) == as.character(orderTable[i,1]))
lost_genes = c(lost_genes,row)
}
gained_genes = c()
for(i in 7726:8725){
row = which(as.character(active_genes[,2])==as.character(orderTable[i,1]))
gained_genes = c(gained_genes,row)
}
pol2_rows_up = c()
for(i in 1:length(gained_genes)){
row = which(rownames(pol2_0)==active_genes[i,1])
pol2_rows_up = c(pol2_rows_up,row)
}
pol2_rows_down = c()
for(i in 1:length(lost_genes)){
row = which(rownames(pol2_0) == active_genes[i,1])
pol2_rows_down = c(pol2_rows_down,row)
}
pol2_24_meta_up = apply(pol2_24[pol2_rows_up,],2,mean,na.rm=TRUE)
print(pol2_24_meta_up[1:5])
print(pol2_24_meta_up[150:155])
print(pol2_24_meta_up[315:320])
pol2_24_meta_down = apply(pol2_24[pol2_rows_down,],2,mean,na.rm=TRUE)
print(pol2_24_meta_down[1:5])
print(pol2_24_meta_down[150:155])
print(pol2_24_meta_down[315:320])
pdf(file='C:\\Users\\rhirsch\\Documents\\rasmc_figures_1-3\\170222_rasmc_gained_vs_lost_24H_meta_750.pdf',width = 10,height = 8)
plot(1:320, pol2_24_meta_up,type='l',col='red',ylim =c(0,2.5),xaxt='n',xlab='',ylab='rpm/bp',main='lost genes 24H (black) vs. gained genes 24H (red)')
lines(1:320, pol2_24_meta_down,type='l',col='black')
axis(1,c(0,60,260,320),c('-3kb','TSS','END','+3kb'))
dev.off()
###########################################
############Gained/Lost Genes Metas 2H#####
###########################################
table2_genes = as.character(rownames(orderTable2))
lost_genes2=c()
for(i in 1:500){
row = which(as.character(active_genes[,2]) == table2_genes[i])
lost_genes2 = c(lost_genes2,row)
}
gained_genes2 = c()
for(i in 7726:8725){
row = which(as.character(active_genes[,2])== table2_genes[i])
gained_genes2 = c(gained_genes2,row)
}
pol2_rows_up2 = c()
for(i in 1:length(gained_genes2)){
row = which(rownames(pol2_0)==active_genes[i,1])
pol2_rows_up2 = c(pol2_rows_up2,row)
}
pol2_rows_down2 = c()
for(i in 1:length(lost_genes2)){
row = which(rownames(pol2_0) == active_genes[i,1])
pol2_rows_down2 = c(pol2_rows_down2,row)
}
pol2_2_meta_up = apply(pol2_2[pol2_rows_up2,],2,mean,na.rm=TRUE)
pol2_2_meta_down = apply(pol2_2[pol2_rows_down2,],2,mean,na.rm=TRUE)
pdf(file='C:\\Users\\rhirsch\\Documents\\rasmc_figures_1-3\\170222_rasmc_gained_vs_lost_2H_meta.pdf',width = 10,height = 8)
plot(1:320, pol2_2_meta_up,type='l',col='red',ylim =c(0,2.5),xaxt='n',xlab='',ylab='rpm/bp',main='lost genes 2H (black) vs. gained genes 2H (red)')
lines(1:320, pol2_2_meta_down,type='l',col='black')
axis(1,c(0,60,260,320),c('-3kb','TSS','END','+3kb'))
dev.off()
|
37ea4ccc462b754f2191118194d7cffb4b481bd4
|
0fd0fb1ada1b966357fe55bfab6540fc7358b62a
|
/man/count_consonants.Rd
|
4ca4dda86f3b6c74710c323dcbf4ba7e610b72f2
|
[
"MIT"
] |
permissive
|
nelsonroque/ruf
|
f1f8598964ade0085ad9ecc77b92b99a4c544a4a
|
dbf86c97ce5ad03e417cd379c47a410c4dbd0566
|
refs/heads/master
| 2021-06-26T18:31:48.057608
| 2021-02-28T14:04:12
| 2021-02-28T14:04:12
| 206,207,103
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 285
|
rd
|
count_consonants.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/count_consonants.R
\name{count_consonants}
\alias{count_consonants}
\title{ruf}
\usage{
count_consonants(str)
}
\arguments{
\item{str}{class: string}
}
\description{
ruf
}
\examples{
count_consonants(str)
}
|
bdfa3f1d3ba92174c9c8200bb1a59c53aa6f4581
|
3c8ba871dfa3dc3673c2fbd0e4838c35e09caf2a
|
/댓글 크롤링.R
|
d032fb9b2c911276c7bed9d3b2898129a2c097b2
|
[] |
no_license
|
JinsaGalbi/webtoon-recommend-system
|
bc580fb61f61802254cf9836eae9a0e7707edf52
|
9983935d1ec26f5723e3859f6c933b08d0886b81
|
refs/heads/master
| 2022-12-15T03:02:37.297373
| 2020-09-08T04:56:15
| 2020-09-08T04:56:15
| 293,414,785
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 7,027
|
r
|
댓글 크롤링.R
|
## 댓글 크롤링
setwd('c:/Users/USER/Desktop/주제분석/웹툰크롤링')
comment_url <- read.csv('comment_url.csv')
library(rvest);library(tidyverse);library(magrittr)
library(RSelenium)
rD <- rsDriver(port = 4445L,browser = 'chrome',chromever = '78.0.3904.70')
remDr <- rD[["client"]]
# comment_final <- data.frame(nickname=NULL,id=NULL,recommend=NULL,unrecommend=NULL,text=NULL,i=NULL)
for(i in 127:length(comment_url$url)){
nickname <- c();id <- c();recommend <- c();unrecommend <- c();date <- c();text <- c()
remDr$navigate(paste0('http://comic.naver.com',comment_url$url[i]))
webElem <- remDr$findElements(using = 'xpath',value = '//*[@id="cbox_module"]/div/div[4]/div[1]/div/ul/li[2]/a') #전체댓글 표시하기
remDr$mouseMoveToLocation(webElement = webElem[[1]]) #마우스 이동
remDr$click() #버튼 클릭
Sys.sleep(0.2)
frontPage <- remDr$getPageSource() #페이지 전체 소스 가져오기
nickname_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_nick') %>% html_text() %>% trimws()
id_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_id') %>% html_text() %>% trimws()
recommend_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_cnt_recomm') %>% html_text() %>% trimws()
unrecommend_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_cnt_unrecomm') %>% html_text() %>% trimws()
date_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_date') %>% html_text() %>% trimws()
text_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_text_wrap') %>% html_text() %>% trimws()
nickname <- c(nickname, nickname_temp)
id <- c(id, id_temp)
recommend <- c(recommend, recommend_temp)
unrecommend <- c(unrecommend, unrecommend_temp)
date <- c(date, date_temp)
text <- c(text, text_temp) # 여기까지 첫페이지
page_flag <- read_html(frontPage[[1]]) %>%
html_nodes(css='.u_cbox_next') %>%
html_attr('href') %>% first()
if(is.na(page_flag)){ # 댓글에 다음페이지가 없는 경우 = 페이지가 1~10까지만 있는 경우
for(j in c(1,3:10)){
webElem <- remDr$findElements(using = 'xpath',value = paste0('//*[@id="cbox_module"]/div/div[7]/div/a[',j,']'))
remDr$mouseMoveToLocation(webElement = webElem[[1]])
remDr$click() #버튼 클릭
Sys.sleep(0.2)
frontPage <- remDr$getPageSource() #페이지 전체 소스 가져오기
nickname_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_nick') %>% html_text() %>% trimws()
id_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_id') %>% html_text() %>% trimws()
recommend_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_cnt_recomm') %>% html_text() %>% trimws()
unrecommend_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_cnt_unrecomm') %>% html_text() %>% trimws()
date_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_date') %>% html_text() %>% trimws()
text_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_text_wrap') %>% html_text() %>% trimws()
nickname <- c(nickname, nickname_temp)
id <- c(id, id_temp)
recommend <- c(recommend, recommend_temp)
unrecommend <- c(unrecommend, unrecommend_temp)
date <- c(date, date_temp)
text <- c(text, text_temp) # 여기까지 2~11페이지
}
}else{ # 댓글에 다음페이지가 있는 경우 = 페이지가 11장 이상
for(k in c(1,3:11)){
webElem <- remDr$findElements(using = 'xpath',value = paste0('//*[@id="cbox_module"]/div/div[7]/div/a[',k,']'))
remDr$mouseMoveToLocation(webElement = webElem[[1]])
remDr$click() #버튼 클릭
Sys.sleep(0.2)
frontPage <- remDr$getPageSource() #페이지 전체 소스 가져오기
nickname_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_nick') %>% html_text() %>% trimws()
id_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_id') %>% html_text() %>% trimws()
recommend_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_cnt_recomm') %>% html_text() %>% trimws()
unrecommend_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_cnt_unrecomm') %>% html_text() %>% trimws()
date_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_date') %>% html_text() %>% trimws()
text_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_text_wrap') %>% html_text() %>% trimws()
nickname <- c(nickname, nickname_temp)
id <- c(id, id_temp)
recommend <- c(recommend, recommend_temp)
unrecommend <- c(unrecommend, unrecommend_temp)
date <- c(date, date_temp)
text <- c(text, text_temp) # 여기까지 2~11페이지
}
for(l in rep(3:12,10000)){
webElem <- remDr$findElements(using = 'xpath',value = paste0('//*[@id="cbox_module"]/div/div[7]/div/a[',l,']'))
if(length(webElem)>=1){
remDr$mouseMoveToLocation(webElement = webElem[[1]])
remDr$click() #버튼 클릭
Sys.sleep(0.2)
frontPage <- remDr$getPageSource() #페이지 전체 소스 가져오기
nickname_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_nick') %>% html_text() %>% trimws()
id_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_id') %>% html_text() %>% trimws()
recommend_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_cnt_recomm') %>% html_text() %>% trimws()
unrecommend_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_cnt_unrecomm') %>% html_text() %>% trimws()
date_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_date') %>% html_text() %>% trimws()
text_temp <- read_html(frontPage[[1]]) %>% html_nodes('.u_cbox_text_wrap') %>% html_text() %>% trimws()
nickname <- c(nickname, nickname_temp)
id <- c(id, id_temp)
recommend <- c(recommend, recommend_temp)
unrecommend <- c(unrecommend, unrecommend_temp)
date <- c(date, date_temp)
text <- c(text, text_temp)
}else{break}}
}
instant <- data.frame(i=i,title=rep(comment_url$title[i],length(id)),nickname=nickname,id=id,text=text) %>% filter(text!='클린봇이 이용자 보호를 위해 숨긴 댓글입니다.') %>% bind_cols(data.frame(recommend=recommend,unrecommend=unrecommend))
comment_final %<>% bind_rows(instant)
}
# write.csv(comment_final,'comment1.csv', row.names = F)
# write.csv(comment_final,'comment2.csv', row.names = F)
# write.csv(comment_final,'comment3.csv', row.names = F)
# write.csv(comment_final,'comment4.csv', row.names = F)
# write.csv(comment_final,'comment5.csv', row.names = F)
# write.csv(comment_final,'comment6.csv', row.names = F)
# write.csv(comment_final,'comment7.csv', row.names = F)
# write.csv(comment_final,'comment8.csv', row.names = F)
# write.csv(comment_final,'comment9.csv', row.names = F)
|
f463c9693a65af962d8d8bba38aa927377cde04a
|
474c866367bda19aa16c88961e34dde81a17786f
|
/housing-market/new-home-sales/new-home-sales.R
|
c2ced9c14db2a14738d5b9ae464dfb4dbc0af4f1
|
[] |
no_license
|
davidallen02/economic-market-commentary
|
fefde0a6e46a5f92eac8dfe9969d004769a34f2e
|
2eef6f4d7fbbff1db39c3323bb7246f6263dff3d
|
refs/heads/master
| 2023-06-08T10:36:40.946661
| 2021-06-22T21:04:09
| 2021-06-22T21:04:09
| 281,976,227
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,320
|
r
|
new-home-sales.R
|
library(magrittr)
lay <- pamngr::set_layout(11)
title <- pamngr::set_title("New Home Sales")
new_home_sales <- pamngr::run_and_load("new-home-sales", "new-home-sales") +
ggplot2::theme(legend.position = "none", plot.caption = ggplot2::element_blank())
northeast <- pamngr::run_and_load("new-home-sales", "new-home-sales-northeast") +
ggplot2::theme(plot.caption = ggplot2::element_blank())
midwest <- pamngr::run_and_load("new-home-sales", "new-home-sales-midwest") +
ggplot2::theme(plot.caption = ggplot2::element_blank())
south <- pamngr::run_and_load("new-home-sales", "new-home-sales-south") +
ggplot2::theme(plot.caption = ggplot2::element_blank())
west <- pamngr::run_and_load("new-home-sales", "new-home-sales-west") +
ggplot2::theme(plot.caption = ggplot2::element_blank())
foo <- gridExtra::grid.arrange(grobs = list(title,
new_home_sales,
northeast,
midwest,
south,
west),
layout_matrix = lay)
ggplot2::ggsave("./housing-market/new-home-sales/new-home-sales.png",
plot = foo, width = 10, height = 6.75, units = "in")
|
2528244c1ab6f5a52508444bafeab1cf52eeada3
|
2d12c1d9fff0c57acc87fb1fcbd9d189854c3d79
|
/programs/ggplottrain.R
|
7427ba1d880cf60c2d7ff4663c157c2283d67d28
|
[
"MIT"
] |
permissive
|
lfthwjx/DataAnalytics
|
8cbab6bf7006d46f6e8ab34bbc86d7907d9ed177
|
68aaf7a936a5a7599e2e2039f6fe26b0a90c14e9
|
refs/heads/master
| 2021-01-13T12:35:18.255673
| 2016-11-01T21:52:37
| 2016-11-01T21:52:37
| 72,579,217
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 585
|
r
|
ggplottrain.R
|
library(ggplot2)
data("diamonds")
p <- ggplot(data = diamonds, mapping = aes(x = carat, y = price, color = cut))
summary(p)
p
p + layer(geom = "point", stat = "identity", position = "identity")
#p + geom_point()
p <- ggplot(data = diamonds, mapping = aes(x = carat))
summary(p)
p
p2 <- p + layer(geom = "bar",stat = "bin")
summary(p2)
p<-ggplot(data=mpg,mapping = aes(x=cty,y=hwy))
p+geom_point()
summary(p)
summary(p+geom_point())
p<-ggplot(data=mpg,aes(x=cty,y=hwy),colours = factor(mpg$year))
p+geom_point()
p+geom_point()+stat_smooth()
factor(mpg$year)
|
2588cb153675ee6087016b010a4cfa5fb084ffe1
|
49e30ad80df564f40b163938f0fd1c00658ff1da
|
/Fig_1B/limma_main_Fig1B.R
|
22fcba4ca58f78a26bcfa550642040b240d5dec5
|
[] |
no_license
|
wasimaftab/IMS_Shotgun
|
ca583f16aa72198ef4ce174c69de7196d528cbc3
|
32bb13f21df969d9562b966f1e44ff9436d3e159
|
refs/heads/master
| 2021-08-18T02:48:43.082139
| 2020-12-27T11:46:04
| 2020-12-27T11:46:04
| 237,218,411
| 1
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 9,580
|
r
|
limma_main_Fig1B.R
|
################################################################################
## This code is to reproduce the data and image of Figure 1B
## Since the missing values are imputed randomly (from normal distribution),
## One can notice minute changes in numbers associated with fold change and p-values
## across multiple runs of the code. However, the major patterns in the data do not change.
## Author: Wasim Aftab
################################################################################
cat('\014')
rm(list = ls())
## Installing Bioconductor packages
if (!requireNamespace("BiocManager", quietly = TRUE))
install.packages("BiocManager")
list.of.packages <- c("limma", "qvalue")
new.packages <- list.of.packages[!(list.of.packages %in% installed.packages()[,"Package"])]
if(length(new.packages)) BiocManager::install(new.packages)
## Installing CRAN packages
list.of.packages <- c("dplyr", "stringr", "MASS", "matlab", "plotly", "htmlwidgets", "rstudioapi", "webshot", "matrixStats")
new.packages <- list.of.packages[!(list.of.packages %in% installed.packages()[,"Package"])]
if(length(new.packages)) install.packages(new.packages)
if (is.null(webshot:::find_phantom())){
webshot::install_phantomjs()
}
library(dplyr)
library(stringr)
library(MASS)
library(matlab)
library(plotly)
library(limma)
library(qvalue)
library(htmlwidgets)
library(rstudioapi)
##Chdir to source dir
path <- rstudioapi::getActiveDocumentContext()$path
Encoding(path) <- "UTF-8"
setwd(dirname(path))
cur_dir <- getwd()
source("limma_helper_functions_Fig1B.R")
##Load the proteingroups file
myFilePath <- "proteinGroups_180416.txt"
proteingroups <-
as.data.frame(read.table(myFilePath, header = TRUE, sep = "\t"))
###Kill the code if proteingroups does not contain crap columns#######################################
temp <-
select(
proteingroups,
matches("(Reverse|Potential.contaminant|Only.identified.by.site)")
)
if (!nrow(temp) * ncol(temp)) {
stop("File error, It does not contain crap...enter another file with crap")
}
######################################################################################################
##Display data to faciliate choice of treatment and control
temp <- select(proteingroups, matches("(ibaq|lfq)"))
print(names(temp))
# #remove "+" identifeid rows from proteingroups##################################################
idx <- NULL
# temp <- as.matrix(select(proteingroups, matches("(Only.identified.by.site|Reverse|Potential.contaminant)")))
temp <- select(proteingroups, matches("(Only.identified.by.site|Reverse|Potential.contaminant)"))
for (i in 1:ncol(temp)){
index <- which(unlist(!is.na(match(temp[,i], "+"))))
idx <- union(idx, index)
}
proteingroups <- proteingroups[-idx, ] # removing indexed rows
# ################################################################################################
#
#Extrat Uniprot and gene symbols
Uniprot <- character(length = nrow(proteingroups))
Symbol <- character(length = nrow(proteingroups))
ProteinNames <- proteingroups$Protein.names
for (i in 1:nrow(proteingroups)) {
temp <- as.character(proteingroups$Fasta.headers[i])
splits <- unlist(strsplit(temp, '\\|'))
Uniprot[i] <- splits[2]
splits <- unlist(str_match(splits[3], "GN=(.*?) PE="))
Symbol[i] <- splits[2]
}
#Extract required data for Limma
treatment <-
readline('Enter treatment name(case insensitive) as it appeared in the iBAQ/LFQ column= ')
control <-
readline('Enter control name(case insensitive) as it appeared in the iBAQ/LFQ column= ')
ibaq <- readinteger_binary('Enter 1 for iBAQ or 0 for LFQ= ')
if (ibaq) {
temp <-
select(proteingroups, matches(paste('^.*', "ibaq", '.*$', sep = '')))
treatment_reps <- data_sanity_check(temp, 'treatment', treatment)
control_reps <- select(temp, matches(control))
control_reps <- data_sanity_check(temp, 'control', control)
data <-
cbind(treatment_reps,
control_reps,
select(proteingroups, matches("^id$")),
Uniprot,
Symbol,
ProteinNames)
} else {
temp <-
select(proteingroups, matches(paste('^.*', "lfq", '.*$', sep = '')))
treatment_reps <- select(temp, matches(treatment))
treatment_reps <- data_sanity_check(temp, 'treatment', treatment)
control_reps <- select(temp, matches(control))
control_reps <- data_sanity_check(temp, 'control', control)
data <-
cbind(treatment_reps,
control_reps,
select(proteingroups, matches("^id$")),
Uniprot,
Symbol,
ProteinNames)
}
print(names(data))
rep_treats <-
readinteger("Enter the number of treatment replicates=")
rep_conts <- readinteger("Enter the number of control replicates=")
FC_Cutoff <- readfloat("Enter the fold change cut off=")
## Find out Blank rows, i.e. proteins with all zeros in treatment and in control, see followig example
# iBAQ.Mrpl40_1 iBAQ.Mrpl40_2 iBAQ.Mrpl40_3 iBAQ.Kgd4_1 iBAQ.Kgd4_2 iBAQ.Kgd4_3 id Uniprot Symbol
#-------------------------------------------------------------------------------------------------
# 0 0 0 0 0 0 84 Q02888 INA17
temp <-
as.matrix(rowSums(apply(data[, 1:(rep_treats + rep_conts)], 2, as.numeric)))
idx <- which(temp == 0)
if (length(idx)) {
data <- data[-idx,] # removing blank rows
}
# ## Find out outliers, i.e. proteins with all zeros in treatment and all/some values in control, see followig example
# ## Uniprot Symbol treat_1 treat_2 treat_3 contrl_1 contrl_2 contrl_3
# ##-----------------------------------------------------------------------------------
# ## P25554 SGF29 0 0 0 2810900 0 0
# temp <- as.matrix(rowSums(data[,1:rep_treats]))
temp <-
as.matrix(rowSums(apply(data[, 1:rep_treats], 2, as.numeric)))
idx <- which(temp == 0)
if (length(idx)) {
outliers <- data[idx,]
filename_outliers <- "Exclusive_proteins_WT"
# paste("Outliers_treatment_", treatment, "_", control, sep = "")
data <- data[-idx, ] # removing indexed rows
}
# ## Find out outliers, i.e. proteins with all zeros in control and all/some values in treatment, see followig example
# iBAQ.Mrpl40_1 iBAQ.Mrpl40_2 iBAQ.Mrpl40_3 iBAQ.Kgd4_1 iBAQ.Kgd4_2 iBAQ.Kgd4_3 id Uniprot Symbol
##-----------------------------------------------------------------------------------------------------
# 662810 505600 559130 0 0 0 79 P38845 CRP1
# temp <-
# as.matrix(rowSums(data[, (rep_treats + 1):(rep_conts + rep_treats)]))
temp <-
as.matrix(rowSums(apply(data[, (rep_treats + 1):(rep_conts + rep_treats)], 2, as.numeric)))
idx <- which(temp == 0)
if (length(idx)) {
outliers_control <- data[idx,]
filename_outliers_control <- "Exclusive_proteins_AROM"
# paste("Outliers_control_", treatment, "_", control, sep = "")
data <- data[-idx, ] # removing indexed rows
}
#Impute data
data_limma <- log2(as.matrix(data[c(1:(rep_treats + rep_conts))]))
data_limma[is.infinite(data_limma)] <- NA
nan_idx <- which(is.na(data_limma))
fit <- fitdistr(c(na.exclude(data_limma)), "normal")
mu <- as.double(fit$estimate[1])
sigma <- as.double(fit$estimate[2])
sigma_cutoff <- 6
new_width_cutoff <- 0.3
downshift <- 1.8
width <- sigma_cutoff * sigma
new_width <- width * new_width_cutoff
new_sigma <- new_width / sigma_cutoff
new_mean <- mu - downshift * sigma
imputed_vals_my = rnorm(length(nan_idx), new_mean, new_sigma)
data_limma[nan_idx] <- imputed_vals_my
##Limma main code
design <-
model.matrix( ~ factor(c(rep(2, rep_treats), rep(1, rep_conts))))
colnames(design) <- c("Intercept", "Diff")
res.eb <- eb.fit(data_limma, design, data$Symbol)
Sig_FC_idx <-
union(which(res.eb$logFC < (-FC_Cutoff)), which(res.eb$logFC > FC_Cutoff))
Sig_Pval_mod_idx <- which(res.eb$p.mod < 0.05)
Sig_Pval_ord_idx <- which(res.eb$p.ord < 0.05)
Sig_mod_idx <- intersect(Sig_FC_idx, Sig_Pval_mod_idx)
Sig_ord_idx <- intersect(Sig_FC_idx, Sig_Pval_ord_idx)
categ_Ord <- rep(c("Not Significant"), times = length(data$Symbol))
categ_Mod <- categ_Ord
categ_Mod[Sig_mod_idx] <- "Significant"
categ_Ord[Sig_ord_idx] <- "Significant"
dat <-
cbind(
res.eb,
categ_Ord,
categ_Mod,
NegLogPvalMod = (-log10(res.eb$p.mod)),
NegLogPvalOrd = (-log10(res.eb$p.ord))
)
dat <- select(dat, -matches("^gene$"))
# dat <- select(dat, -matches("^Uniprot$"))
##Save the data file
final_data <-
cbind(data$Uniprot,
data$ProteinNames,
data$Symbol,
data[,1:(rep_treats+rep_conts)],
dat)
colnames(final_data)[1] <- c("Uniprot")
colnames(final_data)[2] <- c("ProteinNames")
colnames(final_data)[3] <- c("Symbol")
filename_final_data <-
readline('Enter a filename for final data= ')
##Create plotly object and save plot as html
filename_mod <-
readline('Enter a filename for limma plot= ')
# filename_ord <-
# readline('Enter a filename for ordinary t-test plot= ')
# display_plotly_figs(final_data, FC_Cutoff, filename_mod, filename_ord)
display_plotly_figs(final_data, FC_Cutoff, filename_mod)
## Write final data
write.table(
final_data,
paste(filename_final_data, '.tsv', sep = ''),
sep = '\t',
row.names = FALSE,
col.names = TRUE
)
## Write outliers in treatment
write.table(
outliers,
paste(filename_outliers, '.tsv', sep = ''),
sep = '\t',
row.names = FALSE,
col.names = TRUE
)
## Write outliers in control
write.table(
outliers_control,
paste(filename_outliers_control, '.tsv', sep = ''),
sep = '\t',
row.names = FALSE,
col.names = TRUE
)
setwd(cur_dir)
|
f9764567b463f5354b76ef002822db1fba0361e3
|
afa997ac0246d2ede8e3fde18dc3ae780fa16d2f
|
/8kyu/litres.r
|
5876ad0a46bca7e20fc663e7ef83f7baabc57e12
|
[] |
no_license
|
supvolume/codewars_solution_in_R
|
20fc2332afd5b33fd3e15ba716a1e54af955e761
|
1c3912e6b7c60ced07a9879405fad43ada5ba24a
|
refs/heads/master
| 2021-11-27T16:58:02.949777
| 2021-11-26T08:48:05
| 2021-11-26T08:48:05
| 169,952,519
| 1
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 99
|
r
|
litres.r
|
# solution for Keep Hydrated! challenge
litres <- function(time) {
return(as.integer(time*0.5))
}
|
d58f1f833c12d8b6a5a0f0b0473bce3d289ce808
|
ab38fcf1040a66038f81e2963d3ab96cc513a4df
|
/man/lpnet.Rd
|
22628f8581c3cd958e40fc6db696c306d5473c33
|
[] |
no_license
|
yukimayuli-gmz/lpnet
|
a1d271864508886939ede89dac3d73d1d9d3a1f1
|
5cbb3af788c1a742e4448450efca1e1344d81f20
|
refs/heads/main
| 2023-04-19T04:50:32.524818
| 2022-02-17T18:20:03
| 2022-02-17T18:20:03
| 343,287,614
| 1
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 2,529
|
rd
|
lpnet.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/lpnet.R
\name{lpnet}
\alias{lpnet}
\title{Consruct a circular network based on a tree by Linear Programming}
\usage{
lpnet(
M,
tree.method = "unj",
lp.package = "Rglpk",
lp.type = "B",
filename = "lpnet.nex",
taxaname = NULL
)
}
\arguments{
\item{M}{the distance matrix for construct tree and network (the matrix should fit the triangle inequality and the diagonal should be 0).}
\item{tree.method}{method for construct the original tree for lp, default is \code{unj}, for unweighted ntighbor joining tree;
\code{nj} for neighbor joining tree; \code{nnet} for symmetry nnet tree; \code{nnetns} for no symmetry nnet tree;
\code{BioNJ} for BioNJ tree.}
\item{lp.package}{which package will used for Linear Programming, default is \code{Rglpk}, for a free R package;
\code{gurobi} for the gurobi package.}
\item{lp.type}{a character vector indicating the types of the objective variables. default is \code{B} for binary;
\code{I} for intrger; \code{C} for continuous; \code{NULL}, for ordinary.}
\item{filename}{a character will be the naxus file's name, default is \code{lpnet.nex}.}
\item{taxaname}{a character set of names for taxa, ordering is consist with original distance matrix \code{M}.}
}
\value{
The LSfit value.
}
\description{
Construct a planner network which has a circular ordering for a distance matrix, and write a nexus file for SplitsTree4.
First construct a tree for the distance matrix.Then use Linear Programming(lp) to change the circular ordering.
The ordering have the biggest sum of quartets for all taxa is the lp net ordering.
Then use Non-negative least squares(nnls) to calculate weights of splits which are consist with the lp net ordering.
Finally, return a LSfit which is the least squares fit between the pairwise distances in the graph and the pairwise distances in the matrix.
And write a nexus file with taxa block and splits block for SplitsTree4 to see the circular network.
}
\examples{
### From Huson and Bryant (2006, Fig 4):
x <- c(14.06, 17.24, 20.5, 23.37, 17.43, 19.18, 18.48, 9.8, 13.06, 15.93, 15.65,
17.4, 16.7, 6.74, 16.87, 16.59, 18.34, 17.64, 17.57, 17.29, 19.04, 18.34,
17.6, 19.35, 21.21, 9.51, 11.37, 13.12)
M <- matrix(0, 8, 8)
M[row(M) > col(M)] <- x
M <- M + t(M)
taxaname <- c("A", "B", "C", "D", "E", "F", "G", "H")
lpnet(M,
tree.method = "nj",
lp.package = "Rglpk",
lp.type = "B",
filename = "example.nex",
taxaname = taxaname)
}
|
21838bc23052db1f85d945de07e536f808b1a525
|
f256fd8f31ea589cae179c155f44b7e4f8de744a
|
/R/check_collinearity.R
|
163bfcc3d401005f4a9d0d19c7e9a87840471383
|
[] |
no_license
|
cran/performance
|
9578b76fd8981e5896d25d036cb8439e8b01c24a
|
c44285dff9936445c56ec8b83feb7ff9cae3fa81
|
refs/heads/master
| 2023-06-08T08:42:02.230184
| 2023-06-02T10:30:02
| 2023-06-02T10:30:02
| 183,289,241
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 21,547
|
r
|
check_collinearity.R
|
#' @title Check for multicollinearity of model terms
#' @name check_collinearity
#'
#' @description
#'
#' `check_collinearity()` checks regression models for
#' multicollinearity by calculating the variance inflation factor (VIF).
#' `multicollinearity()` is an alias for `check_collinearity()`.
#' `check_concurvity()` is a wrapper around `mgcv::concurvity()`, and can be
#' considered as a collinearity check for smooth terms in GAMs. Confidence
#' intervals for VIF and tolerance are based on Marcoulides et al.
#' (2019, Appendix B).
#'
#' @param x A model object (that should at least respond to `vcov()`,
#' and if possible, also to `model.matrix()` - however, it also should
#' work without `model.matrix()`).
#' @param component For models with zero-inflation component, multicollinearity
#' can be checked for the conditional model (count component,
#' `component = "conditional"` or `component = "count"`),
#' zero-inflation component (`component = "zero_inflated"` or
#' `component = "zi"`) or both components (`component = "all"`).
#' Following model-classes are currently supported: `hurdle`,
#' `zeroinfl`, `zerocount`, `MixMod` and `glmmTMB`.
#' @param ci Confidence Interval (CI) level for VIF and tolerance values.
#' @param verbose Toggle off warnings or messages.
#' @param ... Currently not used.
#'
#' @return A data frame with information about name of the model term, the
#' variance inflation factor and associated confidence intervals, the factor
#' by which the standard error is increased due to possible correlation
#' with other terms, and tolerance values (including confidence intervals),
#' where `tolerance = 1/vif`.
#'
#' @section Multicollinearity:
#' Multicollinearity should not be confused with a raw strong correlation
#' between predictors. What matters is the association between one or more
#' predictor variables, *conditional on the other variables in the
#' model*. In a nutshell, multicollinearity means that once you know the
#' effect of one predictor, the value of knowing the other predictor is rather
#' low. Thus, one of the predictors doesn't help much in terms of better
#' understanding the model or predicting the outcome. As a consequence, if
#' multicollinearity is a problem, the model seems to suggest that the
#' predictors in question don't seems to be reliably associated with the
#' outcome (low estimates, high standard errors), although these predictors
#' actually are strongly associated with the outcome, i.e. indeed might have
#' strong effect (_McElreath 2020, chapter 6.1_).
#'
#' Multicollinearity might arise when a third, unobserved variable has a causal
#' effect on each of the two predictors that are associated with the outcome.
#' In such cases, the actual relationship that matters would be the association
#' between the unobserved variable and the outcome.
#'
#' Remember: "Pairwise correlations are not the problem. It is the conditional
#' associations - not correlations - that matter." (_McElreath 2020, p. 169_)
#'
#' @section Interpretation of the Variance Inflation Factor:
#' The variance inflation factor is a measure to analyze the magnitude of
#' multicollinearity of model terms. A VIF less than 5 indicates a low
#' correlation of that predictor with other predictors. A value between 5 and
#' 10 indicates a moderate correlation, while VIF values larger than 10 are a
#' sign for high, not tolerable correlation of model predictors (_James et al.
#' 2013_). The *Increased SE* column in the output indicates how much larger
#' the standard error is due to the association with other predictors
#' conditional on the remaining variables in the model. Note that these
#' thresholds, although commonly used, are also criticized for being too high.
#' _Zuur et al. (2010)_ suggest using lower values, e.g. a VIF of 3 or larger
#' may already no longer be considered as "low".
#'
#' @section Multicollinearity and Interaction Terms:
#' If interaction terms are included in a model, high VIF values are expected.
#' This portion of multicollinearity among the component terms of an
#' interaction is also called "inessential ill-conditioning", which leads to
#' inflated VIF values that are typically seen for models with interaction
#' terms _(Francoeur 2013)_.
#'
#' @section Concurvity for Smooth Terms in Generalized Additive Models:
#' `check_concurvity()` is a wrapper around `mgcv::concurvity()`, and can be
#' considered as a collinearity check for smooth terms in GAMs."Concurvity
#' occurs when some smooth term in a model could be approximated by one or more
#' of the other smooth terms in the model." (see `?mgcv::concurvity`).
#' `check_concurvity()` returns a column named _VIF_, which is the "worst"
#' measure. While `mgcv::concurvity()` range between 0 and 1, the _VIF_ value
#' is `1 / (1 - worst)`, to make interpretation comparable to classical VIF
#' values, i.e. `1` indicates no problems, while higher values indicate
#' increasing lack of identifiability. The _VIF proportion_ column equals the
#' "estimate" column from `mgcv::concurvity()`, ranging from 0 (no problem) to
#' 1 (total lack of identifiability).
#'
#' @references
#'
#' - Francoeur, R. B. (2013). Could Sequential Residual Centering Resolve
#' Low Sensitivity in Moderated Regression? Simulations and Cancer Symptom
#' Clusters. Open Journal of Statistics, 03(06), 24-44.
#'
#' - James, G., Witten, D., Hastie, T., and Tibshirani, R. (eds.). (2013).
#' An introduction to statistical learning: with applications in R. New York:
#' Springer.
#'
#' - Marcoulides, K. M., and Raykov, T. (2019). Evaluation of Variance
#' Inflation Factors in Regression Models Using Latent Variable Modeling
#' Methods. Educational and Psychological Measurement, 79(5), 874–882.
#'
#' - McElreath, R. (2020). Statistical rethinking: A Bayesian course with
#' examples in R and Stan. 2nd edition. Chapman and Hall/CRC.
#'
#' - Vanhove, J. (2019). Collinearity isn't a disease that needs curing.
#' [webpage](https://janhove.github.io/analysis/2019/09/11/collinearity)
#'
#' - Zuur AF, Ieno EN, Elphick CS. A protocol for data exploration to avoid
#' common statistical problems: Data exploration. Methods in Ecology and
#' Evolution (2010) 1:3–14.
#'
#' @family functions to check model assumptions and and assess model quality
#'
#' @note The code to compute the confidence intervals for the VIF and tolerance
#' values was adapted from the Appendix B from the Marcoulides et al. paper.
#' Thus, credits go to these authors the original algorithm. There is also
#' a [`plot()`-method](https://easystats.github.io/see/articles/performance.html)
#' implemented in the \href{https://easystats.github.io/see/}{\pkg{see}-package}.
#'
#' @examples
#' m <- lm(mpg ~ wt + cyl + gear + disp, data = mtcars)
#' check_collinearity(m)
#'
#' @examplesIf require("see")
#' # plot results
#' x <- check_collinearity(m)
#' plot(x)
#' @export
check_collinearity <- function(x, ...) {
UseMethod("check_collinearity")
}
#' @rdname check_collinearity
#' @export
multicollinearity <- check_collinearity
# default ------------------------------
#' @rdname check_collinearity
#' @export
check_collinearity.default <- function(x, ci = 0.95, verbose = TRUE, ...) {
# check for valid input
.is_model_valid(x)
.check_collinearity(x, component = "conditional", ci = ci, verbose = verbose)
}
# methods -------------------------------------------
#' @export
print.check_collinearity <- function(x, ...) {
insight::print_color("# Check for Multicollinearity\n", "blue")
if ("Component" %in% colnames(x)) {
comp <- split(x, x$Component)
for (i in seq_along(comp)) {
cat(paste0("\n* ", comp[[i]]$Component[1], " component:\n"))
.print_collinearity(datawizard::data_remove(comp[[i]], "Component"))
}
} else {
.print_collinearity(x)
}
invisible(x)
}
#' @export
plot.check_collinearity <- function(x, ...) {
insight::check_if_installed("see", "to plot collinearity-check")
NextMethod()
}
.print_collinearity <- function(x) {
vifs <- x$VIF
low_vif <- which(vifs < 5)
mid_vif <- which(vifs >= 5 & vifs < 10)
high_vif <- which(vifs >= 10)
all_vifs <- insight::compact_list(list(low_vif, mid_vif, high_vif))
# if we have no CIs, remove those columns
x <- datawizard::remove_empty_columns(x)
# format table for each "ViF" group - this ensures that CIs are properly formatted
x <- insight::format_table(x)
x <- datawizard::data_rename(
x,
pattern = "SE_factor",
replacement = "Increased SE",
verbose = FALSE
)
if (length(low_vif)) {
cat("\n")
insight::print_color("Low Correlation\n\n", "green")
print.data.frame(x[low_vif, ], row.names = FALSE)
}
if (length(mid_vif)) {
cat("\n")
insight::print_color("Moderate Correlation\n\n", "yellow")
print.data.frame(x[mid_vif, ], row.names = FALSE)
}
if (length(high_vif)) {
cat("\n")
insight::print_color("High Correlation\n\n", "red")
print.data.frame(x[high_vif, ], row.names = FALSE)
}
}
# other classes ----------------------------------
#' @export
check_collinearity.afex_aov <- function(x, verbose = TRUE, ...) {
if (length(attr(x, "within")) == 0L) {
return(check_collinearity(x$lm, verbose = verbose, ...))
}
f <- insight::find_formula(x)[[1]]
f <- Reduce(paste, deparse(f))
f <- sub("\\+\\s*Error\\(.*\\)$", "", f)
f <- stats::as.formula(f)
d <- insight::get_data(x, verbose = FALSE)
is_num <- vapply(d, is.numeric, logical(1))
d[is_num] <- sapply(d[is_num], datawizard::center, simplify = TRUE)
is_fac <- !is_num
contrs <- lapply(is_fac, function(...) stats::contr.sum)[is_fac]
if (verbose) {
insight::format_alert(
"All predictors have been centered (factors with `contr.sum()`, numerics with `scale()`)."
)
}
check_collinearity(suppressWarnings(stats::lm(
formula = f,
data = d,
contrasts = contrs
)))
}
#' @export
check_collinearity.BFBayesFactor <- function(x, verbose = TRUE, ...) {
if (!insight::is_model(x)) {
insight::format_error("Collinearity only applicable to regression models.")
}
f <- insight::find_formula(x)[[1]]
d <- insight::get_data(x, verbose = FALSE)
check_collinearity(stats::lm(f, d))
}
# mfx models -------------------------------
#' @export
check_collinearity.logitor <- function(x, ci = 0.95, verbose = TRUE, ...) {
.check_collinearity(x$fit, component = "conditional", ci = ci, verbose = verbose)
}
#' @export
check_collinearity.logitmfx <- check_collinearity.logitor
#' @export
check_collinearity.probitmfx <- check_collinearity.logitor
#' @export
check_collinearity.poissonirr <- check_collinearity.logitor
#' @export
check_collinearity.poissonmfx <- check_collinearity.logitor
#' @export
check_collinearity.negbinirr <- check_collinearity.logitor
#' @export
check_collinearity.negbinmfx <- check_collinearity.logitor
#' @export
check_collinearity.betaor <- check_collinearity.logitor
#' @export
check_collinearity.betamfx <- check_collinearity.logitor
# zi-models -------------------------------------
#' @rdname check_collinearity
#' @export
check_collinearity.glmmTMB <- function(x,
component = c("all", "conditional", "count", "zi", "zero_inflated"),
ci = 0.95,
verbose = TRUE,
...) {
component <- match.arg(component)
.check_collinearity_zi_model(x, component, ci = ci, verbose = verbose)
}
#' @export
check_collinearity.MixMod <- function(x,
component = c("all", "conditional", "count", "zi", "zero_inflated"),
ci = 0.95,
verbose = TRUE,
...) {
component <- match.arg(component)
.check_collinearity_zi_model(x, component, ci = ci, verbose = verbose)
}
#' @export
check_collinearity.hurdle <- function(x,
component = c("all", "conditional", "count", "zi", "zero_inflated"),
ci = 0.95,
verbose = verbose,
...) {
component <- match.arg(component)
.check_collinearity_zi_model(x, component, ci = ci, verbose = verbose)
}
#' @export
check_collinearity.zeroinfl <- function(x,
component = c("all", "conditional", "count", "zi", "zero_inflated"),
ci = 0.95,
verbose = verbose,
...) {
component <- match.arg(component)
.check_collinearity_zi_model(x, component, ci = ci, verbose = verbose)
}
#' @export
check_collinearity.zerocount <- function(x,
component = c("all", "conditional", "count", "zi", "zero_inflated"),
ci = 0.95,
verbose = verbose,
...) {
component <- match.arg(component)
.check_collinearity_zi_model(x, component, ci = ci, verbose = verbose)
}
# utilities ---------------------------------
.check_collinearity_zi_model <- function(x, component, ci = 0.95, verbose = TRUE) {
if (component == "count") component <- "conditional"
if (component == "zi") component <- "zero_inflated"
mi <- insight::model_info(x, verbose = FALSE)
if (!mi$is_zero_inflated) component <- "conditional"
if (component == "all") {
cond <- .check_collinearity(x, "conditional", ci = ci, verbose = verbose)
zi <- .check_collinearity(x, "zero_inflated", ci = ci, verbose = FALSE)
if (is.null(cond) && is.null(zi)) {
return(NULL)
}
if (is.null(cond)) {
zi$Component <- "zero inflated"
return(zi)
}
if (is.null(zi)) {
cond$Component <- "conditional"
return(cond)
}
# retrieve data for plotting
dat_cond <- attr(cond, "data")
dat_zi <- attr(zi, "data")
ci_cond <- attr(cond, "CI")
ci_zi <- attr(zi, "CI")
# add component
cond$Component <- "conditional"
zi$Component <- "zero inflated"
dat_cond$Component <- "conditional"
dat_zi$Component <- "zero inflated"
ci_cond$Component <- "conditional"
ci_zi$Component <- "zero inflated"
# create final data
dat <- rbind(cond, zi)
attr(dat, "data") <- rbind(dat_cond, dat_zi)
attr(dat, "CI") <- rbind(ci_cond, ci_zi)
dat
} else {
.check_collinearity(x, component, ci = ci, verbose = verbose)
}
}
.check_collinearity <- function(x, component, ci = 0.95, verbose = TRUE) {
v <- insight::get_varcov(x, component = component, verbose = FALSE)
assign <- .term_assignments(x, component, verbose = verbose)
# any assignment found?
if (is.null(assign) || all(is.na(assign))) {
if (verbose) {
insight::format_alert(
sprintf("Could not extract model terms for the %s component of the model.", component)
)
}
return(NULL)
}
# we have rank-deficiency here. remove NA columns from assignment
if (isTRUE(attributes(v)$rank_deficient) && !is.null(attributes(v)$na_columns_index)) {
assign <- assign[-attributes(v)$na_columns_index]
if (isTRUE(verbose)) {
insight::format_alert(
"Model matrix is rank deficient. VIFs may not be sensible."
)
}
}
# check for missing intercept
if (insight::has_intercept(x)) {
v <- v[-1, -1]
assign <- assign[-1]
} else {
if (isTRUE(verbose)) {
insight::format_alert("Model has no intercept. VIFs may not be sensible.")
}
}
f <- insight::find_formula(x)
if (inherits(x, "mixor")) {
terms <- labels(x$terms)
} else {
terms <- labels(stats::terms(f[[component]]))
}
if ("instruments" %in% names(f)) {
terms <- unique(c(terms, labels(stats::terms(f[["instruments"]]))))
}
n.terms <- length(terms)
if (n.terms < 2) {
if (isTRUE(verbose)) {
insight::format_alert(
sprintf("Not enough model terms in the %s part of the model to check for multicollinearity.", component)
)
}
return(NULL)
}
R <- stats::cov2cor(v)
detR <- det(R)
result <- vector("numeric")
na_terms <- vector("numeric")
for (term in 1:n.terms) {
subs <- which(assign == term)
if (length(subs)) {
result <- c(
result,
det(as.matrix(R[subs, subs])) * det(as.matrix(R[-subs, -subs])) / detR
)
} else {
na_terms <- c(na_terms, term)
}
}
# any terms to remove, due to rank deficiency?
if (length(na_terms)) {
terms <- terms[-na_terms]
}
# check for interactions, VIF might be inflated...
if (!is.null(insight::find_interactions(x)) && any(result > 10) && isTRUE(verbose)) {
insight::format_alert(
"Model has interaction terms. VIFs might be inflated.",
"You may check multicollinearity among predictors of a model without interaction terms."
)
}
# CIs, see Appendix B 10.1177/0013164418817803
r <- 1 - (1 / result)
n <- insight::n_obs(x)
p <- insight::n_parameters(x)
# check if CIs are requested
if (!is.null(ci) && !is.na(ci) && is.numeric(ci)) {
ci_lvl <- (1 + ci) / 2
logis_r <- stats::qlogis(r) # see Raykov & Marcoulides (2011, ch. 7) for details.
se <- sqrt((1 - r^2)^2 * (n - p - 1)^2 / ((n^2 - 1) * (n + 3)))
se_log <- se / (r * (1 - r))
ci_log_lo <- logis_r - stats::qnorm(ci_lvl) * se_log
ci_log_up <- logis_r + stats::qnorm(ci_lvl) * se_log
ci_lo <- stats::plogis(ci_log_lo)
ci_up <- stats::plogis(ci_log_up)
} else {
ci_lo <- ci_up <- NA
}
out <- insight::text_remove_backticks(
data.frame(
Term = terms,
VIF = result,
VIF_CI_low = 1 / (1 - ci_lo),
VIF_CI_high = 1 / (1 - ci_up),
SE_factor = sqrt(result),
Tolerance = 1 / result,
Tolerance_CI_low = 1 - ci_up,
Tolerance_CI_high = 1 - ci_lo,
stringsAsFactors = FALSE
),
column = "Term"
)
attr(out, "ci") <- ci
attr(out, "data") <- insight::text_remove_backticks(
data.frame(
Term = terms,
VIF = result,
SE_factor = sqrt(result),
stringsAsFactors = FALSE
),
column = "Term"
)
attr(out, "CI") <- data.frame(
VIF_CI_low = 1 / (1 - ci_lo),
VIF_CI_high = 1 / (1 - ci_up),
Tolerance_CI_low = 1 - ci_up,
Tolerance_CI_high = 1 - ci_lo,
stringsAsFactors = FALSE
)
class(out) <- c("check_collinearity", "see_check_collinearity", "data.frame")
out
}
.term_assignments <- function(x, component, verbose = TRUE) {
tryCatch(
{
if (inherits(x, c("hurdle", "zeroinfl", "zerocount"))) {
assign <- switch(component,
conditional = attr(insight::get_modelmatrix(x, model = "count"), "assign"),
zero_inflated = attr(insight::get_modelmatrix(x, model = "zero"), "assign")
)
} else if (inherits(x, "glmmTMB")) {
assign <- switch(component,
conditional = attr(insight::get_modelmatrix(x), "assign"),
zero_inflated = .zi_term_assignment(x, component, verbose = verbose)
)
} else if (inherits(x, "MixMod")) {
assign <- switch(component,
conditional = attr(insight::get_modelmatrix(x, type = "fixed"), "assign"),
zero_inflated = attr(insight::get_modelmatrix(x, type = "zi_fixed"), "assign")
)
} else {
assign <- attr(insight::get_modelmatrix(x), "assign")
}
if (is.null(assign)) {
assign <- .find_term_assignment(x, component, verbose = verbose)
}
assign
},
error = function(e) {
.find_term_assignment(x, component, verbose = verbose)
}
)
}
.find_term_assignment <- function(x, component, verbose = TRUE) {
pred <- insight::find_predictors(x)[[component]]
if (is.null(pred)) {
return(NULL)
}
dat <- insight::get_data(x, verbose = FALSE)[, pred, drop = FALSE]
parms <- unlist(lapply(seq_along(pred), function(i) {
p <- pred[i]
if (is.factor(dat[[p]])) {
ps <- paste0(p, levels(dat[[p]]))
names(ps)[seq_along(ps)] <- i
ps
} else {
names(p) <- i
p
}
}))
if (insight::is_gam_model(x)) {
model_params <- as.vector(unlist(insight::find_parameters(x)[c(component, "smooth_terms")]))
} else {
model_params <- insight::find_parameters(x)[[component]]
}
as.numeric(names(parms)[match(
insight::clean_names(model_params),
parms
)])
}
.zi_term_assignment <- function(x, component = "zero_inflated", verbose = TRUE) {
tryCatch(
{
rhs <- insight::find_formula(x)[[component]]
d <- insight::get_data(x, verbose = FALSE)
attr(insight::get_modelmatrix(rhs, data = d), "assign")
},
error = function(e) {
NULL
}
)
}
|
f016d45af91e02c9c900332fe90a551519a3522a
|
5db2dac679963587ac50ad850ea3a2ccb508465a
|
/phd-scripts/R/zeta_plot.R
|
575804f6dcd92b8239de7789259cd8f1a01b7996
|
[
"MIT"
] |
permissive
|
softloud/simeta
|
be88fe336eeee9610086823adce839493781c0ef
|
2a7e979077c57812a7d29c3e23e8c00080e1cb03
|
refs/heads/master
| 2023-04-16T23:27:16.936986
| 2023-03-25T11:49:23
| 2023-03-25T11:49:23
| 200,359,586
| 2
| 2
|
NOASSERTION
| 2020-01-28T09:55:16
| 2019-08-03T09:56:12
|
HTML
|
UTF-8
|
R
| false
| false
| 1,525
|
r
|
zeta_plot.R
|
#' plot the distribution of the proportion allocated to the intervention group
#'
#' @family vis_tools
#' @family reporting Functions and tools for reporting simulation results.
#'
#' @export
zeta_plot <- function(mu, epsilon) {
# check numeric args
neet::assert_neet(mu, "numeric")
neet::assert_neet(epsilon, "numeric")
# check args are [0,1]
assertthat::assert_that(
mu > 0 & mu < 1,
msg = "mu must be a value from [0,1].")
assertthat::assert_that(
epsilon > 0 & epsilon < 1,
msg = "epsilon must be a value from [0,1].")
# calculate parameters
par <- beta_par(mu, epsilon)
# return plot of beta distribution with parameters
tibble(x = c(0, 1)) %>%
ggplot(aes(x = x)) +
geom_rect(xmin = mu - epsilon, xmax = mu + epsilon, ymin = 0, ymax = Inf, alpha = 0.2) +
geom_vline(xintercept = mu, linetype = "dashed", alpha = 0.8) +
stat_function(fun = dbeta,
linetype = "dotted",
args = list(shape1 = par$alpha, shape2 = par$beta)) +
labs(title = str_wrap("Distribution of expected proportion of intervention cohort"
),
x = TeX("$\\zeta$"),
y = NULL,
caption = str_wrap(paste0(
"We assume a beta distribution with expected centre ",
mu,
" and 90% of values falling within ",
epsilon,
"; i.e, within the interval [",
mu - epsilon,
",",
mu + epsilon,
"]"), width = 70)) +
theme(axis.text.y = element_blank(), axis.ticks.y = element_blank())
}
|
f9ca07dcd19963980785a8e19f5ac735565975cd
|
5e67544bd977d277ea24050d1aafa1c9bed9cf86
|
/2-analysis/old/adoption.R
|
574966b90e8e35102c8ab9efc418e202775850ff
|
[] |
no_license
|
balachia/Currency
|
8d08a1c11a6472e7b019c641afb64ad90c7e1b7b
|
ff46e4b042eb176cb7787ba524c52d21303cd5ce
|
refs/heads/master
| 2021-01-17T04:46:31.851629
| 2016-07-14T22:45:06
| 2016-07-14T22:45:06
| 15,240,711
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 8,994
|
r
|
adoption.R
|
# let's run an adoption model viz coxph
# What do we need to do here?
# Create a data set containing events by user/currency
# split spells on ...
# every day with observed friend or self returns?
library(rms)
library(survival)
library(data.table)
library(ffbase)
library(reshape2)
library(texreg)
library(parallel)
rm(list=ls())
qcuts <- function(x, qs) {
cut(x, unique(quantile(x, qs, na.rm=TRUE)), labels=FALSE, include.lowest=TRUE)
}
collapse.to.spells <- function(dt, t.var, grp.vars, sort.vars, coll.vars) {
dt.out <- data.table(dt)
dt.out <- dt[do.call(order, dt[,c(grp.vars,sort.vars), with=FALSE])]
dt.out[, nspell := FALSE]
lapply(coll.vars, function (x) {
dt.out[, nspell := nspell | (get(x) != c(NA, get(x)[1:(.N-1)])),by=grp.vars]
0
})
dt.out[is.na(nspell), nspell := TRUE]
dt.out[,spellid := cumsum(nspell)]
dt.out[, start := min(get(t.var)) - 1, by=spellid]
dt.out[, stop := max(get(t.var)), by=spellid]
# print(dt.out)
# return(dt.out)
dt.out <- dt.out[ dt.out[,.I[.N], by=spellid][,V1]]
dt.out
}
poor.cem <- function(dt, keys, snames=NULL, qnames=NULL, bkeys=keys) {
kdt <- dt[,c(keys,snames),with=FALSE]
setkeyv(dt,keys)
setkeyv(kdt,keys)
print(kdt)
# print(names(dt))
cat('making quantiles\n')
lapply(qnames, function (x) {
cat('processing ',x,'\n')
kdt[,paste0(x,'_q') := qcuts(dt[,get(x)], seq(0,1,0.1))]
NULL
})
allnames <- c(snames, paste0(qnames,'_q'))
cat('making groups\n')
kdt[,grp := .GRP, by=allnames]
ngrps <- kdt[,max(grp)]
cat('# Groups:', ngrps, '\n')
cat('balance statistics\n')
cat('# observations\n')
dat <- kdt[,.N,by=grp][,N]
print(summary(dat))
print(quantile(dat,seq(0,1,0.1), na.rm=TRUE))
cat('\n')
for(key in bkeys) {
cat(key,':\n')
# dat <- kdt[,dim(.SD[,.N,by=key])[1],by=grp][,V1]
dat <- kdt[,length(unique(get(key))),by=grp][,V1]
# try a parallel version...
# dats <- mclapply(split(1:ngrps, factor(1:ngrps %% (64 * 2))),
# mc.cores=64, mc.preschedule=FALSE,
# function (subgrps) {
# sub.dat <- kdt[grp %in% subgrps, length(unique(key)), by=grp][,V1]
# sub.dat
# })
# dat <- rbindlist(dats)
# print(dats)
# print(dat)
print(summary(dat))
print(quantile(dat,seq(0,1,0.1), na.rm=TRUE))
cat('# == 1 ::', sum(dat==1), '(', sum(dat==1) / length(dat), '%)\n')
cat('\n')
}
kdt[,c(keys,'grp'), with=FALSE]
}
if (grepl('.*stanford\\.edu',Sys.info()[['nodename']])) {
DATA.DIR <- '/archive/gsb/vashevko/forex/'
OUT.DIR <- '~/2YP/writing/'
} else {
DATA.DIR <- '~/Data/forex/'
OUT.DIR <- '~/Dropbox/forex Project/writing/'
}
setwd(DATA.DIR)
# load in data
aus <- readRDS('./Rds/active-user-quantiles.Rds')
dt <- readRDS('./Rds/day.stats-0-1samp.Rds')
fpt <- readRDS('./Rds/forexposition.Rds')
ffdfns <- load.ffdf('./ffdb/sbc')
ffd <- ffdfns$ffd
# put days in fpt stuff
fpt[,openday := floor((opendate / 86400000.0) + 0.125)]
fpt[,cp := paste0(currency1,currency2)]
# get currency use frequency
cps <- fpt[, list(cpN=.N), by=cp]
cps <- cps[order(-cps$cpN)]
cps[,rank:=.I]
# merge in currency frequency
# why?
setkey(cps,cp)
setkey(fpt,cp)
fpt <- merge(fpt,cps,all.x=TRUE)
uids <- aus[Npd_q %in% 2:5 &
med_ob_q %in% 2:5 &
dpnlpd_q %in% 2:5 &
netdep_q %in% 2:5,user_id]
# reduce users under consideration
c.dt <- dt[user_id %in% uids]
# pull out all adoption events
g.min.day <- 1199145600000 / 86400000
all.adopt.es <- fpt[,list(adopt = min(openday)), by=list(user_id,cp)]
# throw out each user's first day of currency
# all.adopt.es <- all.adopt.es[order(user_id,adopt)]
all.adopt.es[, valid := adopt > min(adopt), by=user_id]
all.adopt.es <- all.adopt.es[(valid)]
all.adopt.es <- all.adopt.es[adopt > g.min.day]
all.adopt.es[,valid := NULL]
# can we pull in the whole sbc database? only need results of single day...
idx.1d <- ffwhich(ffd,gap==1)
sbc.1d <- as.data.table(as.data.frame(ffd[idx.1d,]))
print(object.size(sbc.1d), units='auto')
# create time since last trade
c.dt <- c.dt[order(user_id,day)]
c.dt[, bopen := as.integer(opened_today > 0)]
c.dt[, cumopen := cumsum(bopen) - bopen, by=user_id]
c.dt[, dlapse := 1:.N, by=list(user_id,cumopen)]
#c.dt[, dlapse := as.factor(dlapse)]
c.dt[, c('imputed','bopen','cumopen') := NULL]
# we should insert user's own stats here...
sbc.1d <- sbc.1d[, c('gap','imputed','dpnl_sum','dpnl_mean','dpnl_pos','dpnl_neg') := NULL]
sbc.e1d <- sbc.1d[type == 'ego']
sbc.a1d <- sbc.1d[type == 'alter']
# make 10 day lag for ego
# make 14 day lag for alters
sbc.a14 <- rbindlist(lapply(0:13, function (x) {
res <- data.table(sbc.a1d)
res[, src := day]
res[, day := day + x]
res
}))
sbc.e2 <- rbindlist(lapply(0:1, function (x) {
res <- data.table(sbc.e1d)
res[, src := day]
res[, day := day + x]
res
}))
sbc.e10 <- rbindlist(lapply(2:9, function (x) {
res <- data.table(sbc.e1d)
res[, src := day]
res[, day := day + x]
res
}))
# now collapse the fuckers
sbc.a14 <- sbc.a14[, list(ntotal.alt = sum(ntotal),
npos.alt = sum(npos),
nneg.alt = sum(nneg)
), by=list(user_id, cp, day)]
sbc.e2 <- sbc.e2[, list(ntotal.e2 = sum(ntotal),
npos.e2 = sum(npos),
nneg.e2 = sum(nneg)
), by=list(user_id, day)]
sbc.e10 <- sbc.e10[, list(ntotal.e10 = sum(ntotal),
npos.e10 = sum(npos),
nneg.e10 = sum(nneg)
), by=list(user_id, day)]
# sort the bastard
setkey(sbc.a14, user_id, day)
setkey(sbc.e2, user_id, day)
setkey(sbc.e10, user_id, day)
setkey(c.dt,user_id, day)
c.dt <- merge(c.dt, sbc.e2, all.x=TRUE)
c.dt <- merge(c.dt, sbc.e10, all.x=TRUE)
c.dt[is.na(ntotal.e2), c('ntotal.e2', 'npos.e2', 'nneg.e2') := 0]
c.dt[is.na(ntotal.e10), c('ntotal.e10', 'npos.e10', 'nneg.e10') := 0]
# set up the users/currencies under observation
cp.set <- all.adopt.es[,unique(cp)]
# cp.set <- cp.set[1:4]
# cp.set <- c('EURUSD','USDCZK')
# cp.set <- cps[(rank - 1) %% 10 == 0, cp]
# what do we need in the spell split?
res <- mclapply(cp.set, mc.cores=60, mc.preschedule=FALSE,
function (ccp) {
print(ccp)
# adopt.es <- fpt[cp == ccp, list(adopt=min(openday)), by=user_id]
adopt.es <- all.adopt.es[cp == ccp]
cp.rank <- cps[cp==ccp, rank]
setkey(adopt.es,user_id)
cp.dt <- merge(c.dt, adopt.es, all.x=TRUE)
cp.dt <- cp.dt[is.na(adopt) | day <= adopt]
cp.dt[, badopt := ifelse(day==adopt,1,0)]
cp.dt[is.na(badopt), badopt := 0]
cp.dt[, cp := ccp]
cp.dt[, adopt := NULL]
setkey(cp.dt, user_id, day)
cp.dt <- merge(cp.dt, sbc.a14[cp == ccp, !'cp', with=FALSE], all.x=TRUE)
cp.dt[is.na(ntotal.alt), c('ntotal.alt', 'npos.alt', 'nneg.alt') := 0]
cp.dt[,rank := cp.rank]
#
cp.dt
})
if(!file.exists('dta/adopts-m10.dta')) {
cat('writing to stata\n')
library(foreign)
all.adopts <- rbindlist(res)
print(object.size(all.adopts), units='auto')
print(dim(all.adopts))
cat('starting write\n')
write.dta(all.adopts, 'dta/adopts-m10.dta')
cat('done with write\n')
}
# put day groups back in
res <- lapply(res, function(x) {
x[, dgrp := cumsum(opened_today>0) - (opened_today>0), by=user_id]
x
})
res2 <- mclapply(res, mc.cores=60, mc.preschedule=FALSE,
function (cdt) {
colldt <- collapse.to.spells(cdt, 'dlapse', c('user_id','cp'), 'day',
coll.vars=c('ntotal.e2','ntotal.e10','ntotal.alt','dgrp'))
colldt
})
# saveRDS(res2, 'Rds/adopt.events.m10.Rds')
saveRDS(res2, 'Rds/adopt.events.Rds')
stop()
stop()
lapply(res, function (cpdt) {
print(m.res <- clogit(badopt ~ ntotal.alt + strata(dlapse), data=cpdt))
0
})
stop()
stopifnot(FALSE)
# assign user group
aus <- aus[user_id %in% uids]
aus[,ugrp := .GRP, by=list(nfriends_q,naccts_q,med_ob_q,netdep_q,Npd_q,dpnlpd_q)]
# aus[,ugrp := .GRP, by=list(nfriends_q,naccts_q,med_ob_q,Npd_q,dpnlpd_q)]
aus[,ugrpN := .N, by=ugrp]
summary(aus[,.N,by=ugrp])
quantile(aus[,.N,by=ugrp][,N], seq(0,1,0.1))
# drop single user groups
# uids <- aus[ugrpN > 1, user_id]
uids <- aus[,user_id]
# pull out user-restricted observations
c.dt <- dt[user_id %in% uids]
c.fpt <- fpt[user_id %in% uids]
# preprocess c.dt stuff
c.dt[,bopen:=as.integer(opened_today>0)]
# merge in user statistics
setkey(c.dt,user_id)
setkey(aus,user_id)
c.dt <- merge(c.dt,aus)
|
4fd2c388a19e7e3a237edacd3c7114b175d47838
|
1f10c23dcc48836a6f731a5dd3a1ba1279efd343
|
/R_scripts/3_fit_models.R
|
b3fa668dcf28667726fd47226066597dd9797aae
|
[
"CC-BY-2.0"
] |
permissive
|
LeanaGooriah/ISAR_analysis
|
fcab2596754ca3ba7a75b7499ec749effaf94b2c
|
2827a402db297464ace5c29222f7bf735bcc007c
|
refs/heads/master
| 2021-07-15T02:44:43.837907
| 2020-05-31T22:24:44
| 2020-05-31T22:24:44
| 147,322,370
| 6
| 1
| null | null | null | null |
UTF-8
|
R
| false
| false
| 692
|
r
|
3_fit_models.R
|
library(tidyr)
library(purrr)
library(dplyr)
library(broom)
dat1 <- read.csv("ISAR_DATA/diversity_indices/allstudies_allscales_allindices.csv", stringsAsFactors = F)
names(dat1)
by_index <- dat1 %>%
group_by(Study, Scale, index) %>%
nest()
by_index2 <- by_index %>%
mutate(model = map(data, ~ lm(log(value) ~ log(Area), data = .)),
Intercept = map_dbl(model, ~coef(.x)[1]),
Slope = map_dbl(model, ~coef(.x)[2]),
glance = map(model, glance)
) %>%
unnest(glance)
names(by_index2)
by_index2a <- by_index2 %>%
select(Study:r.squared, -model, -data, p.value)
write.csv(by_index2a, "ISAR_DATA/results/Table_2.csv", row.names = FALSE)
|
7dc5a5425d33c2fd427c9d304f72075f9e33c339
|
4d2c711b308db9eeefd2faf1f13d4c5db3ab70b1
|
/plot1.R
|
b0e4ca019c5a70ec1fa13f01068e731cabd3b833
|
[] |
no_license
|
kimhale/ExData_Plotting1
|
1ad185478180f4b969fb7093d92616ac2d9fcf48
|
53014e056de9524bc888444ff2368a9d9d271aa1
|
refs/heads/master
| 2020-03-23T02:11:07.152949
| 2018-07-14T22:41:53
| 2018-07-14T22:41:53
| 140,961,037
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,480
|
r
|
plot1.R
|
# Set working directory
setwd("/Users/kimberlyhale/Documents/Coursera/DataScienceCert/C4_EDA/ExData_Plotting1")
# Load packages
library(data.table)
# Download file if data doesn't exist
if (!file.exists("data/household_power_consumption.txt")){
dir.create("data")
fileURL <- "https://d396qusza40orc.cloudfront.net/exdata%2Fdata%2Fhousehold_power_consumption.zip"
download.file(fileURL, destfile = "data/HouseholdPower.zip", method = "curl")
#Unzip file
zipF<- "data/HouseholdPower.zip"
outDir<-"data"
unzip(zipF,exdir=outDir)
}
# Read data into datatable
dt <- fread("data/household_power_consumption.txt", na.strings = "?")
# Check that data is as expected
head(dt)
summary(dt)
# Turn Date into Date class and filter to just Feb 01-02 2007
dt$Date <- as.Date(dt$Date, format = "%d/%m/%Y")
dt <- dt[(dt$Date >= as.Date("2007-02-01") & dt$Date <= as.Date("2007-02-02")), ]
# Turn Date/Time into Date/Time instead of character
dt$DateTime <- paste(dt$Date, dt$Time, sep = " ")
dt$DateTime <- as.POSIXct(dt$DateTime, format = "%Y-%m-%d %H:%M:%S")
# Check that data is as expected
dt$DateTime[2] - dt$DateTime[1]
head(dt$DateTime)
# Create a histogram, change title, color, x-axis title
# Save it to a PNG file, width of 480 pixels and height of 480 pixels, named plot1.png
png(filename = 'plot1.png', width = 480, height = 480)
hist(dt$Global_active_power, col = "red", main = "Global Active Power",
xlab = "Global Active Power (kilowatts)")
dev.off()
|
3dd29b0e9348838f15d725dff2952d512be0a7cb
|
7e1d6c1822045ee656a6a41c063631760466add3
|
/man/appendVerticalTab.Rd
|
113dd66ba17e586cb312a4e3dc34c7214f741149
|
[
"MIT"
] |
permissive
|
jcheng5/shinyWidgets
|
c784a3c9e4da76c0c7f23e8362fe647044beb6d2
|
f17f91f6c2e38ee8d7c6be6484ccb474ebef6417
|
refs/heads/master
| 2020-04-29T18:20:46.195001
| 2019-03-18T16:10:00
| 2019-03-18T16:56:59
| 176,321,196
| 3
| 0
|
NOASSERTION
| 2019-03-18T15:59:35
| 2019-03-18T15:59:35
| null |
UTF-8
|
R
| false
| true
| 1,350
|
rd
|
appendVerticalTab.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/vertical-tab.R
\name{appendVerticalTab}
\alias{appendVerticalTab}
\alias{removeVerticalTab}
\alias{reorderVerticalTabs}
\title{Mutate Vertical Tabset Panel}
\usage{
appendVerticalTab(inputId, tab,
session = shiny::getDefaultReactiveDomain())
removeVerticalTab(inputId, index,
session = shiny::getDefaultReactiveDomain())
reorderVerticalTabs(inputId, newOrder,
session = shiny::getDefaultReactiveDomain())
}
\arguments{
\item{inputId}{The id of the \code{verticalTabsetPanel} object.}
\item{tab}{The verticalTab to append.}
\item{session}{The \code{session} object passed to function given to \code{shinyServer.}}
\item{index}{The index of the the tab to remove.}
\item{newOrder}{The new index order.}
}
\description{
Mutate Vertical Tabset Panel
}
\examples{
if (interactive()) {
library(shiny)
library(shinyWidgets)
ui <- fluidPage(
verticalTabsetPanel(
verticalTabPanel("blaa","foo"),
verticalTabPanel("yarp","bar"),
id="hippi"
)
)
server <- function(input, output, session) {
appendVerticalTab("hippi", verticalTabPanel("bipi","long"))
removeVerticalTab("hippi", 1)
appendVerticalTab("hippi", verticalTabPanel("howdy","fair"))
reorderVerticalTabs("hippi", c(3,2,1))
}
# Run the application
shinyApp(ui = ui, server = server)
}
}
|
f88419bd06c98c170a02b78d9937aff2a86fcd8d
|
83d7c5a5d018752961787e7ddaee30e005ab36aa
|
/man/bbx_pad.Rd
|
fab8f72bc6f95f6a186b0037aae8412c55bc4df3
|
[
"MIT"
] |
permissive
|
dmi3kno/bbx
|
09f9997463987144b0c3a541ae96215598a96405
|
c805c50bc1ebbab04c035b80e4c5f835979d6f91
|
refs/heads/master
| 2022-06-22T03:44:02.044482
| 2020-05-07T08:22:31
| 2020-05-07T08:22:31
| 261,409,960
| 2
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 1,028
|
rd
|
bbx_pad.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/transform.R
\name{bbx_pad_width}
\alias{bbx_pad_width}
\alias{bbx_pad_height}
\title{Functions for padding bbx
These functions can "pad" (increase size of) bbx}
\usage{
bbx_pad_width(bbx, n = 1, word = NULL, side = "both")
bbx_pad_height(bbx, n = 1, word = NULL, side = "both")
}
\arguments{
\item{bbx}{character vector of bounding boxes to pad}
\item{n}{integer number of pixels to add. If a `word` is provided, `n` is interpreted as number of characters.}
\item{word}{optional character vector of words contained in bbxes}
\item{side}{"left", "right" (for `bbx_pad_width()`), "up", "down" (for `bbx_pad_height()`) or "both" which side to pad}
}
\value{
a vector of validated bbxes
}
\description{
Functions for padding bbx
These functions can "pad" (increase size of) bbx
}
\examples{
bbx_pad_width("5 5 10 20", word="There")
bbx_pad_width("5 5 10 20", 1)
bbx_pad_height("5 5 10 20", word="There/nbe/ndragons")
bbx_pad_height("5 5 10 20", 1)
}
|
9a362c140c0f61805bf401a7c3c87046f4f53bd8
|
c9a0e70c007ab6f1f6495dbaef44258e18959f17
|
/climate_comp_plots_AGUV_GUV_IVT_AGU2020poster.R
|
a7eff26545bf9cc74904cd038cc65112062b70eb
|
[
"MIT"
] |
permissive
|
LizCarter492/MW_P_STG
|
e51e10ef872222fbc644f6ca88836e48dd13cb8c
|
06701a69c29201eb717638baecd4ec1c0e25e88c
|
refs/heads/main
| 2023-03-06T18:34:00.191059
| 2021-02-20T01:11:07
| 2021-02-20T01:11:07
| 340,173,894
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 19,232
|
r
|
climate_comp_plots_AGUV_GUV_IVT_AGU2020poster.R
|
library(lubridate)
library(ncdf4)
library(fields)
library(maps)
library(maptools)
library(raster)
library(data.table)
library(RColorBrewer)
library(viridis)
library(colorRamps)
#setwd("D:/DELL_PHD/Box Sync/AFRI_Aexam/R scripts")
source("filled.contour3.R")
source("filled.legend.R")
in_path<-("D:/DELL_PHD/E/NASH_subseasonal_prediction/")
out_path<-("H:/MW_seasonal_warm_season_precip_forecast/")
######################################################
#Variable names:######################################
#gu: geostrophic u wind component
#gv: geostrophic v wind component
#agu: ageostrophic u wind component
#agv: ageostrophic v wind component
#p: precipitation
#mdf: moisture flux divergence
#ivt: integrated vapor transport
#q: specific humidity
#o: omega (surface lifting index)
#z: geopotential height
#h: sensible heat flux
#
#ltm prefix indicates long-term monthly mean
#######################################################
#######################################################
#(u,v); (gu, gv); (agu, agv): load total, gesotrophic, and ageostrophic wind####
gUV<-nc_open(paste(in_path,"geos_uv_1000_600mB.nc", sep=""))
gu<-ncvar_get(gUV, varid="u_g")
gv<-ncvar_get(gUV, varid="v_g")
nc_close(gUV)
unc<-nc_open(paste(in_path,"uwnd.mon.mean.nc", sep=""))
u<-ncvar_get(unc, varid="uwnd")
u<-u[,,1:5,]
nc_close(unc)
vnc<-nc_open(paste(in_path,"vwnd.mon.mean.nc", sep=""))
v<-ncvar_get(vnc, varid="vwnd")
v<-v[,,1:5,]
nc_close(vnc)
agu<-u-gu
agu<-apply(agu[,,1:4,1:840], c(1,2,4), mean)
agv<-v-gv
agv<-apply(agv[,,1:4,1:840], c(1,2,4), mean)
gu<-apply(gu[,,1:4,1:840], c(1,2,4), mean)
gv<-apply(gv[,,1:4,1:840], c(1,2,4), mean)
u<-apply(u[,,1:4,1:840], c(1,2,4), mean)
v<-apply(v[,,1:4,1:840], c(1,2,4), mean)
ltmagu<-ltmagv<-ltmgu<-ltmgv<-ltmu<-ltmv<-
array(NA, dim=c(dim(agu)[1:2],12))
mthi<-rep(1:12, length(1948:2017))
yeari<-rep(1948:2017, each=12)
for(j in 1:12){
ltmagu[,,j]<-apply(agu[,,which(mthi==j)],c(1,2), mean)
ltmagv[,,j]<-apply(agv[,,which(mthi==j)],c(1,2), mean)
ltmgu[,,j]<-apply(gu[,,which(mthi==j)],c(1,2), mean)
ltmgv[,,j]<-apply(gv[,,which(mthi==j)],c(1,2), mean)
ltmu[,,j]<-apply(u[,,which(mthi==j)],c(1,2), mean)
ltmv[,,j]<-apply(v[,,which(mthi==j)],c(1,2), mean)
}
#(P): load GPCC precipitation####
setwd(in_path)
P.nc<-nc_open("precip.comb.v2018to2016-v6monitorafter.total.nc")
plat<-ncvar_get(P.nc, "lat")
plon<-ncvar_get(P.nc, "lon")
ptime<-ncvar_get(P.nc, "time")
ptime<-as.POSIXct("1800-01-01 00:00")+as.difftime(ptime,units="days")
ptime<-as.Date(ptime, origin=as.Date("1800-01-01 00:00:0.0"))
p<-ncvar_get(P.nc, "precip")
p<-p[,,which(year(ptime)>=1948 & year(ptime)<2018)]
nc_close(P.nc)
#(z): load geopotential height#####
setwd(in_path)
Z.nc<-nc_open("hgt.mon.mean.nc")
lat<-ncvar_get(Z.nc, "lat")
lon<-ncvar_get(Z.nc, "lon")
lev<-ncvar_get(Z.nc, "level")
z<-ncvar_get(Z.nc, "hgt")
time<-ncvar_get(Z.nc, "time")
time<-as.POSIXct("1800-01-01 00:00")+as.difftime(time,units="hours")
time<-as.Date(time, origin=as.Date("1800-01-01 00:00:0.0"))
z<-z[,,,year(time)<2018]
nc_close(Z.nc)
ltmz<-array(NA, dim=c(dim(z)[1:3],12))
mthi<-rep(1:12, length(1948:2017))
yeari<-rep(1948:2017, each=12)
for(j in 1:12){
ltmz[,,,j]<-apply(z[,,,which(mthi==j)],c(1,2,3), mean)
}
#(MFD): load moisture flux divergence####
MFD<-array(NA, dim=c(144, 73, 12*length(1948:2018)))
mthi<-rep(1:12, length(1948:2017))
yeari<-rep(1948:2017, each=12)
setwd(paste(in_path,"NCAR_NCEP_Daily_moisture_budget/", sep=""))
for(g in (1948:2018)){
mfd.nc <- nc_open(paste("MFD_daily_", g, ".nc", sep=""))
time<-ncvar_get(mfd.nc, varid="time")
time<-as.POSIXct("1800-01-01 00:00")+as.difftime(time,units="hours")
time<-as.Date(time, origin=as.Date("1800-01-01 00:00:0.0"))
mfd<-ncvar_get(mfd.nc, "Div_UQVQ")
nc_close(mfd.nc)
mth<-month(time)
for(j in 1:12){
MFD[,,which(mthi==j & yeari==g)]<-apply(mfd[,,mth==j], c(1,2), mean)
}
remove(mfd)
}
ltmMFD<-array(NA, dim=c(dim(MFD)[1:2],12))
for(j in 1:12){
ltmMFD[,,j]<-apply(MFD[,,which(mthi==j)],c(1,2), mean)
}
#(UQ, VQ, IVT): load integrated vapor transport####
UQ<-VQ<-IVT<-array(NA, dim=c(144, 73, 12*length(1948:2017)))
mthi<-rep(1:12, length(1948:2017))
yeari<-rep(1948:2017, each=12)
setwd(paste(in_path,"NCAR_NCEP_Daily_moisture_budget/", sep=""))
for(g in (1948:2017)){
VINT.nc <- nc_open(paste("VINT_daily_", g, ".nc", sep=""))
time<-ncvar_get(VINT.nc, varid="time")
time<-as.POSIXct("1800-01-01 00:00")+as.difftime(time,units="hours")
time<-as.Date(time, origin=as.Date("1800-01-01 00:00:0.0"))
uq<-ncvar_get(VINT.nc, "UQ")
vq<-ncvar_get(VINT.nc, "VQ")
nc_close(VINT.nc)
mth<-month(time)
for(j in 1:12){
UQ[,,which(mthi==j & yeari==g)]<-apply(uq[,,mth==j], c(1,2), mean)
VQ[,,which(mthi==j & yeari==g)]<-apply(vq[,,mth==j], c(1,2), mean)
}
remove(uq); remove(vq); remove(time); remove(mth)
}
IVT<-sqrt(UQ^2 + VQ^2)
ltmUQ<-ltmVQ<-ltmIVT<-array(NA, dim=c(dim(MFD)[1:2],12))
for(j in 1:12){
ltmUQ[,,j]<-apply(UQ[,,which(mthi==j)],c(1,2), mean)
ltmVQ[,,j]<-apply(VQ[,,which(mthi==j)],c(1,2), mean)
ltmIVT[,,j]<-apply(IVT[,,which(mthi==j)],c(1,2), mean)
}
#(q): load specific humidity####
setwd(in_path)
Q.nc<-nc_open("shum.mon.mean.nc")
lat<-ncvar_get(Q.nc, "lat")
lon<-ncvar_get(Q.nc, "lon")
q<-ncvar_get(Q.nc, "shum")
time<-ncvar_get(Q.nc, "time")
time<-as.POSIXct("1800-01-01 00:00")+as.difftime(time,units="hours")
time<-as.Date(time, origin=as.Date("1800-01-01 00:00:0.0"))
q<-apply(q[,,1:4,year(time)<2018], c(1,2,4), mean)
ltmq<-array(NA, dim=c(dim(q)[1:2],12))
for(j in 1:12){
ltmq[,,j]<-apply(q[,,which(mthi==j)],c(1,2), mean)
}
#(o): load omega####
setwd(in_path)
O.nc<-nc_open("omega.mon.mean.nc")
lat<-ncvar_get(O.nc, "lat")
lon<-ncvar_get(O.nc, "lon")
o<-ncvar_get(O.nc, "omega")
time<-ncvar_get(O.nc, "time")
time<-as.POSIXct("1800-01-01 00:00")+as.difftime(time,units="hours")
time<-as.Date(time, origin=as.Date("1800-01-01 00:00:0.0"))
o<-apply(o[,,1:4,year(time)<2018], c(1,2,4), mean)
ltmo<-array(NA, dim=c(dim(o)[1:2],12))
for(j in 1:12){
ltmo[,,j]<-apply(o[,,which(mthi==j)],c(1,2), mean)
}
#
#####################################################################################
#set map graphical parameters####
#map dimensions
lat.min <- 10#min(lat)
lat.max <- 75#max(lat)
lon.min <- -132#min(lon-360)
lon.max <- max(lon-360)
lat.min.ar <- lat.min+1
lat.max.ar <- lat.max-1
lon.min.ar <- lon.min+1
lon.max.ar <- lon.max-1
lat_final <- rev(lat)
lon_final <- lon - 360
keep1 <- which(lon_final>=lon.min & lon_final<=lon.max)
keep2 <- which(lat_final>=lat.min & lat_final<=lat.max)
lon_all <- rep(lon_final,length(lat_final))
lat_all <- sort(rep(lat_final,length(lon_final)))
month_name<-format(ISOdate(2004,1:12,1),"%B")
mth<-rep(1:12, length(1948:2017))
yr<-rep(1948:2017, each=12)
#image size
hgt<-500
wdt<-1600
p_c_df<-rbind(c(1),c(2),c(3), c(4), c(5), c(6), c(7), c(8), c(9), c(10), c(11), c(12))
lat.min <- 10#min(lat)
lat.max <- 75#max(lat)
lon.min <- -132#min(lon-360)
lon.max <- max(lon-360)
lat.min.ar <- lat.min+1
lat.max.ar <- lat.max-1
lon.min.ar <- lon.min+1
lon.max.ar <- lon.max-1
lat_final <- rev(lat)
lon_final <- lon - 360
keep1 <- which(lon_final>=lon.min & lon_final<=lon.max)
keep2 <- which(lat_final>=lat.min & lat_final<=lat.max)
lon_all <- rep(lon_final,length(lat_final))
lat_all <- sort(rep(lat_final,length(lon_final)))
month_name<-format(ISOdate(2004,1:12,1),"%B")
mth<-rep(1:12, length(1948:2017))
yr<-rep(1948:2017, each=12)
#image size
hgt<-1500 #4 #
wdt<-3200 #12.8
for(m in 1:12){
m_c<-p_c_df[m,]
#mfd
ltmUQ_sub<-apply(ltmUQ[,,m_c],c(1,2),mean)
ltmUQ_sub<-t(ltmUQ_sub)
ltmUQ_sub<-ltmUQ_sub[seq(dim(ltmUQ_sub)[1],1),]
ltmVQ_sub<-apply(ltmVQ[,,m_c],c(1,2),mean)
ltmVQ_sub<-t(ltmVQ_sub)
ltmVQ_sub<-ltmVQ_sub[seq(dim(ltmVQ_sub)[1],1),]
ltmmfd_sub<-apply(ltmMFD[,,m_c], c(1,2), mean)*30
ltmmfd_sub<-t(ltmmfd_sub)
ltmmfd_sub<-ltmmfd_sub[seq(dim(ltmmfd_sub)[1],1),]
#geo
ltmz850_sub<-apply(ltmz[,,3,m_c],c(1,2),mean)
ltmz850_sub<-t(ltmz850_sub)
ltmz850_sub<-ltmz850_sub[seq(dim(ltmz850_sub)[1],1),]
gu_sub<-apply(gu[,,which(mth %in% (m_c))],c(1,2), mean)
gu_sub<-t(gu_sub)
gu_sub<-gu_sub[seq(dim(gu_sub)[1],1),]
gv_sub<-apply(gv[,,which(mth %in% (m_c))],c(1,2), mean)
gv_sub<-t(gv_sub)
gv_sub<-gv_sub[seq(dim(gv_sub)[1],1),]
#ageo
ltmo_sub<-apply(o[,,which (mth %in% m_c)], c(1,2), mean)
ltmo_sub<-t(ltmo_sub)
ltmo_sub<-ltmo_sub[seq(dim(ltmo_sub)[1],1),]
agu_sub<-apply(agu[,,which(mth %in% (m_c))],c(1,2), mean)
agu_sub<-t(agu_sub)
agu_sub<-agu_sub[seq(dim(agu_sub)[1],1),]
agv_sub<-apply(agv[,,which(mth %in% (m_c))],c(1,2), mean)
agv_sub<-t(agv_sub)
agv_sub<-agv_sub[seq(dim(agv_sub)[1],1),]
png(paste("H:/Research/MW_ClimateChange/images/climatological maps/", month_name[m_c[1]],"_", month_name[m_c[2]],"climatologyREVISED.png",sep=""),
res=300,
#pointsize=15,
type="cairo",
width=wdt, height=hgt,
)
par(mfrow=c(1,2))
#geo
uua<-as.vector(t(gu_sub))
vva<-as.vector(t(gv_sub))
uu<-uua
vv<-vva
speed <- sqrt(uu*uu+ vv*vv)
speeda <- sqrt(uua*uua + vva*vva)
uv <-which(speed>speeda & (1:length(uu))%in%seq(2,length(uu),by=4))
uvall<-which(speed<=speeda & (1:length(uu))%in%seq(2,length(uu),by=4))
mylevel<-c(1200, seq(1400, 1600, by=5), 1700)
mycol<-topo.colors(length(mylevel)+1)
mnm<-c("January", "February", "March", "April", "May", "June", "July", "August", "September", "October", "November", "December")[m]
filled.contour3(lon_final,lat_final,t(ltmz850_sub),
xlim=c(lon.min,lon.max),
ylim=c(lat.min,lat.max),
level=mylevel,
col=mycol,
#frame.plot=FALSE,
main=title(mnm,cex.main=3),
plot.axes={
axis(1, labels=FALSE, tick=FALSE);
axis(2, labels=FALSE, tick=FALSE);
if(m %in% 4:9){
arrow.plot(a1=lon_all[uv],a2=lat_all[uv],u=uua[uv],v=vva[uv],arrow.ex=0.5,length=0.06,xpd=FALSE,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=3, col="gray40");
arrow.plot(a1=lon_all[uvall],a2=lat_all[uvall],u=uua[uvall],v=vva[uvall],arrow.ex=0.5,length=0.06,xpd=FALSE,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=3, col="gray40");
}else{
arrow.plot(a1=lon_all[uv],a2=lat_all[uv],u=uua[uv],v=vva[uv],arrow.ex=0.5,length=0.06,xpd=FALSE,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=3, col="gray40");
arrow.plot(a1=lon_all[uvall],a2=lat_all[uvall],u=uua[uvall],v=vva[uvall],arrow.ex=0.5,length=0.06,xpd=FALSE,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=3, col="gray40");
}
if(m_c[1]<5){
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1480,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1420,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1540,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub),lwd=1,lty=1, levels=c(1440,1500,1520,1540, 1560),
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max),add=TRUE);
}else if(m_c[1]<6) {
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1480,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1420,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1540,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1500,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1440,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub),lwd=1,lty=1, levels=c(1440,1500,1520,1540, 1560),
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max),add=TRUE);
}else{
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1480,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1420,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1540,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1510,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1460,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub), levels=1560,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=4,lty=1, add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmz850_sub),lwd=1,lty=1, levels=c(1440,1500,1520,1540),
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), add=TRUE);
}
map("world",add=T, lwd=3);
}
)
# #ageo
# uua<-as.vector(t(agu_sub))
# vva<-as.vector(t(agv_sub))
# uu<-uua
# uua[uu>quantile(uu, 0.99, na.rm=T)]<-NA#quantile(uu, 0.95, na.rm=T)
# #uua[uu<quantile(uu, 0.01, na.rm=T)]<-NA#quantile(uu, 0.05, na.rm=T)
# vv<-vva
# vva[vva>quantile(vv, 0.99, na.rm=T)]<-NA#quantile(vva, 0.95, na.rm=T)
# #vva[vv<quantile(vv, 0.05, na.rm=T)]<-NA#quantile(vv, 0.05, na.rm=T)
# #uua<-ifelse(uua>=0, sqrt(abs(uua)), -1*sqrt(abs(uua)))
# #vva<-ifelse(vva>=0, sqrt(abs(vva)), -1*sqrt(abs(vva)))
# speed <- sqrt(uua*uua+ vva*vva)
# speeda <- sqrt(uua*uua + vva*vva)
# uv <-which(speed>speeda & (1:length(uu))%in%seq(1,length(uu),by=2))
# uvall<-which(speed<=speeda & (1:length(uu))%in%seq(1,length(uu),by=2))
#
# mylevel<-c(-0.2, seq(-0.02, 0.02, by=0.001), 0.5)
# mycol<-cm.colors(length(mylevel)+1)
#
# filled.contour3(lon_final,lat_final,t(ltmo_sub),
# xlim=c(lon.min,lon.max),
# ylim=c(lat.min,lat.max),
# level=mylevel,
# col=mycol,
# main=title(mnm,cex.main=3),
# plot.axes={
# axis(1, labels=FALSE, tick=FALSE);
# axis(2, labels=FALSE, tick=FALSE);
# arrow.plot(a1=lon_all[uv],a2=lat_all[uv],u=uua[uv],v=vva[uv],arrow.ex=0.5,length=0.06,xpd=FALSE,
# xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=3, col="gray40");
# arrow.plot(a1=lon_all[uvall],a2=lat_all[uvall],u=uua[uvall],v=vva[uvall],arrow.ex=0.5,length=0.06,xpd=FALSE,
# xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=3, col="gray40");
# contour(x=lon_final, y=lat_final, z=t(ltmo_sub), levels=c(0.008),
# xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=2,lty=1, col="black",add=TRUE);
# contour(x=lon_final, y=lat_final, z=t(ltmo_sub), levels= c(-0.008),lty=2,
# xlim=c(lon.min,lon.max),ylim=c(lat.min+5,lat.max), lwd=2, col="black",add=TRUE);
#
# map("world",add=T, lwd=3);
# }
#
# )
#mfd
library(colorRamps)
mylevel<-c(-7, seq(-3, 3, by=0.1), 10)
mycol<-rev(matlab.like(length(mylevel)+1))
uua<-as.vector(t(ltmUQ_sub))
vva<-as.vector(t(ltmVQ_sub))
#uua<-ifelse(uua>=0, sqrt(abs(uua)), -1*sqrt(abs(uua)))
#vva<-ifelse(vva>=0, sqrt(abs(vva)), -1*sqrt(abs(vva)))
uvall<-which((1:length(uua))%in%seq(1,length(uua),by=3))
filled.contour3(lon_final,lat_final,t(ltmmfd_sub),
xlim=c(lon.min,lon.max),
ylim=c(lat.min,lat.max),
level=mylevel,
col=mycol,
#frame.plot=FALSE,
key.title=title(main=paste(j, k, sep=" "),cex=2),
plot.axes={
axis(1, labels=FALSE, tick=FALSE);
axis(2, labels=FALSE, tick=FALSE);
arrow.plot(a1=lon_all[uvall],a2=lat_all[uvall],u=uua[uvall],v=vva[uvall],arrow.ex=0.4,length=0.06,xpd=FALSE,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=3, col="grey40");
contour(x=lon_final, y=lat_final, z=t(ltmmfd_sub), levels= c(-1.5),lty=2,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=3, col="black",add=TRUE);
contour(x=lon_final, y=lat_final, z=t(ltmmfd_sub), levels= c(1.5),lty=1,
xlim=c(lon.min,lon.max),ylim=c(lat.min,lat.max), lwd=3, col="black",add=TRUE);
map("world",add=T, lwd=3);
}
)
dev.off()
}
|
fdd28f2ca586a74ae6acf749263d5fe263120948
|
b75b290b2dd161e4c850858ecf7abee486bdede8
|
/man/calc_cgp.Rd
|
7589849018715d0f17fccd34f387a42f88fc3405
|
[] |
no_license
|
rabare/mapvizieR
|
7d7bffb8691f6045e678d822f9e461e748038cd3
|
1a344ec1376ee41e85dcda0407f8bc63cfb95a82
|
refs/heads/master
| 2020-12-25T16:14:33.669385
| 2015-07-14T21:24:27
| 2015-07-14T21:24:27
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 930
|
rd
|
calc_cgp.Rd
|
% Generated by roxygen2 (4.1.1): do not edit by hand
% Please edit documentation in R/cgp_prep.R
\name{calc_cgp}
\alias{calc_cgp}
\title{calc_cgp}
\usage{
calc_cgp(measurementscale, grade, growth_window, baseline_avg_rit = NA,
ending_avg_rit = NA, sch_growth_study = sch_growth_norms_2012,
calc_for = c(1:99))
}
\arguments{
\item{measurementscale}{MAP subject}
\item{grade}{baseline/starting grad for the group of students}
\item{growth_window}{desired growth window for targets (fall/spring, spring/spring, fall/fall)}
\item{baseline_avg_rit}{the baseline mean rit for the group of students}
\item{ending_avg_rit}{the baseline mean rit for the group of students}
\item{sch_growth_study}{a school growth study to use. default is sch_growth_norms_2012}
\item{calc_for}{vector of cgp targets to calculate for.}
}
\value{
a named list - targets, and results
}
\description{
calculates both cgp targets and cgp results.
}
|
6c7f8064efec1b49029180adb0f75d1594ae282d
|
2171f3867747a89929b1ad396afb855b09896309
|
/man/yan_2013.Rd
|
832f7396b51bb9b54906ccc61faac0d2f994aa83
|
[
"MIT"
] |
permissive
|
aelhossiny/rDeepMAPS
|
4c1a0a0bcdad4bf6240984b9a47da55250c1fa8e
|
ec6b742f9f42dc4f1a40a392ddfe8d3610ab9e63
|
refs/heads/master
| 2023-06-09T20:02:19.866887
| 2021-07-07T02:56:57
| 2021-07-07T02:56:57
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 504
|
rd
|
yan_2013.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/data.R
\docType{data}
\name{yan_2013}
\alias{yan_2013}
\title{Data of yan_2013}
\format{
A list with expression matrix and metadata:
$meta:
\describe{
\item{$expr}{a data frame with 20214 genes (rows) amd 90 cells (columns)}
\item{$meta$Cluster}{fct cell type}
}
}
\source{
\url{https://www.nature.com/articles/nsmb.2660}
}
\usage{
yan_2013
}
\description{
Yan 2013: Human embryo, 7 cell types, 90 cells
}
\keyword{datasets}
|
215ba762b7cc4a8e6ab9bc81badb0841c5fe9406
|
b54ab392f5b31b6d665521d7f32d9b77ccae8728
|
/aprioritestingCosmetics.R
|
9f428d5a243942c07d466365d65feb820d52063c
|
[] |
no_license
|
snehavishwanatha/Apriori_on_Cosmetics_Dataset
|
59cd97bdf5600e5a52b479faf165b7adbf5a88b9
|
dd145ddc4e5c1487fdbc368d31ab864df4d0652c
|
refs/heads/master
| 2020-05-15T21:45:01.909665
| 2019-04-23T07:12:51
| 2019-04-23T07:12:51
| 182,508,244
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 3,272
|
r
|
aprioritestingCosmetics.R
|
library(arules)
library(arulesViz)
library(RColorBrewer)
f=file.choose()
f1=read.csv(f)
#summarizing data
data<-f1
#data=na.omit(f1)
head(data,n=10)
str(data)
summary(data)
#inspecting rules
rules<-apriori(data,parameter=list(supp=0.5,conf=0.8,target="rules"))
rules.sorted <-sort(rules, by="confidence")
rules <- rules.sorted[!is.redundant(rules)]
summary(rules)
inspect(rules)
#item frequency histogram
rules<-apriori(data,parameter=list(supp=0.5,conf=0.8,target="rules"))
itemFrequencyPlot(items(rules),topN=13,col=brewer.pal(8,'Pastel2'),
main='Relative Item Frequency Plot',type="relative",ylab="Item Frequency (Relative)")
#graphical understanding
plot(rules[1:25],method = "graph",cex=0.7,main="Graphical representation for 25 rules")
#reporting the client in layman terms
sink("cosmeticsreport.txt")
suppyes=vector()
supportyes=vector()
suppno=vector()
supportno=vector()
for(q in 1:ncol(data))
{
suppyes[q]=0
suppno[q]=0
}
#suppyes
#suppno
for(i in 1:ncol(data))
{
for(j in 1:nrow(data))
{
if(data[j,i]=="No")
suppno[i]=suppno[i]+1
if(data[j,i]=="Yes")
suppyes[i]=suppyes[i]+1
}
supportyes[i]=suppyes[i]/nrow(data)
supportno[i]=suppno[i]/nrow(data)
if(supportyes[i]>0.5)
print(paste("Probability of the product ",colnames(data[i])," being purchased is ",supportyes[i]*100,"%"))
}
#suppno
#supportno
#suppyes
#supportyes
op=vector()
sop=vector()
for(l in 1:ncol(data[,-1]))
{ s=l+1
for(k in s:ncol(data))
{
if(l!=k)
{
for(opq in 1:4)
op[opq]=0
for(j in 1:nrow(data))
{
if(data[j,l]=="No"&&data[j,k]=="No")
{
op[1]=op[1]+1
}
else if(data[j,l]=="Yes"&&data[j,k]=="No")
{
op[2]=op[2]+1
}
else if(data[j,l]=="No"&&data[j,k]=="Yes")
{
op[3]=op[3]+1
}
else op[4]=op[4]+1
}
sop=op
for(b in 1:4)
{
if(op[b]!=0)
op[b]=op[b]/nrow(data)
}
if(sop[4]!=0&&op[4]>0.5&&suppyes[l]!=0)
{
sop[4]=sop[4]/suppyes[l]
print(paste(" The probability that purchase of ",colnames(data[l])," influences the purchase of ",colnames(data[k])," when placed on the same aisle is"))
print(paste(sop[4]*100,"%"))
}
}
}
}
op=vector()
sop=vector()
for(l in ncol(data[,-1]):2)
{ s=l-1
for(k in s:ncol(data))
{
if(l!=k)
{
for(opq in 1:4)
op[opq]=0
for(j in 1:nrow(data))
{
if(data[j,l]=="No"&&data[j,k]=="No")
{
op[1]=op[1]+1
}
else if(data[j,l]=="Yes"&&data[j,k]=="No")
{
op[2]=op[2]+1
}
else if(data[j,l]=="No"&&data[j,k]=="Yes")
{
op[3]=op[3]+1
}
else op[4]=op[4]+1
}
sop=op
for(b in 1:4)
{
if(op[b]!=0)
op[b]=op[b]/nrow(data)
}
if(sop[4]!=0&&op[4]>0.6&&suppyes[l]!=0)
{
sop[4]=sop[4]/suppyes[l]
print(paste(" The probability that purcahse of ",colnames(data[l])," influences the purchase of ",colnames(data[k])," when placed on the same aisle is"))
print(paste(sop[4]*100," %"))
}
}
}
}
sink()
|
eef2d94954c515dd5df756cec2f12b4c4fd722dc
|
8c07cac2a097b87ab88301826126271165bdd261
|
/november/data/nadieh/Step 1 - Get the top 100 Fantasy authors from Amazon.R
|
8bfe38970bd69a1b6fde013c0a10793e220212aa
|
[] |
no_license
|
Minotaur-zx/datasketches-gh-pages
|
d53dc7e7d3afac9e0a279eb95cb44463c69950f1
|
b63dcef736d67a52da69d70092f5eb25be4e8fb5
|
refs/heads/master
| 2022-12-31T08:06:05.991245
| 2020-10-23T05:35:06
| 2020-10-23T05:35:06
| 306,538,550
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,093
|
r
|
Step 1 - Get the top 100 Fantasy authors from Amazon.R
|
#Get the top 100 fantasy authors from Amazon
#Taken on November 7th, 2016 at 13:24 Amsterdam time
#Web scraper explanation
#https://blog.rstudio.org/2014/11/24/rvest-easy-web-scraping-with-r/
library(rvest)
library(stringr)
library(tidyr)
#https://www.amazon.com/author-rank/Fantasy/books/16190/ref=kar_mr_pg_1?_encoding=UTF8&pg=1
#https://www.amazon.com/author-rank/Fantasy/books/16190/ref=kar_mr_pg_1?_encoding=UTF8&pg=2
numPages <- 10
authorList <- NULL
for(i in 1:numPages) {
url <- paste('https://www.amazon.com/author-rank/Fantasy/books/16190/ref=kar_mr_pg_1?_encoding=UTF8&pg=',i,sep="")
webpage <- read_html(url)
#Get the names of the authors
#Used http://selectorgadget.com/ to figure out how to grab a hold of the author's name
authorListPage <- webpage %>%
html_nodes(".kar_authorName") %>%
html_text()
authorList <- append(authorList, authorListPage)
}#for i
#save the list
authorListDF <- data.frame(rank = 1:length(authorList), author = authorList, stringsAsFactors = F)
write.csv(authorListDF, file="top100FantasyAuthorsAmazon.csv", row.names = F)
|
26abe9f81a64e8c4bcd8f7696660a6d584701f9f
|
a8b27b2a52ed577253ecc30b969ee8f4d5b0dca3
|
/man/opchar_admissable.Rd
|
629134e020523462c65f0786d20e6582a9d67964
|
[] |
no_license
|
mjg211/singlearm
|
51d2a0c3141ef4861a0156e0b29d57c23815100b
|
ad0c23a7780b902b1d5028915b8948d208999ba6
|
refs/heads/master
| 2021-06-03T14:25:29.390692
| 2021-05-04T14:57:48
| 2021-05-04T14:57:48
| 133,361,647
| 0
| 1
| null | null | null | null |
UTF-8
|
R
| false
| true
| 2,486
|
rd
|
opchar_admissable.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/opchar_admissable.R
\name{opchar_admissable}
\alias{opchar_admissable}
\title{Determine the operating characteristics of admissable group sequential
single-arm trial designs for a single binary endpoint}
\usage{
opchar_admissable(des, k, pi, summary = F)
}
\arguments{
\item{des}{An object of class \code{"sa_des_admissable"}, as returned by
\code{des_admissable()}.}
\item{k}{Calculations are performed conditional on the trial stopping in one
of the stages listed in vector \code{k}. Thus, \code{k} should be a vector of
integers, with elements between one and the maximal number of possible stages
in the supplied designs. If left unspecified, it will internally default to
all possible stages.}
\item{pi}{A vector of response probabilities to evaluate operating
characteristics at. This will internally default to be the
\ifelse{html}{\out{<i>π</i><sub>0</sub>}}{\deqn{\pi_0}} and
\ifelse{html}{\out{<i>π</i><sub>1</sub>}}{\deqn{\pi_1}} from the
supplied designs if it is left unspecified.}
\item{summary}{A logical variable indicating whether a summary of the
unction's progress should be printed to the console.}
}
\value{
A list of class \code{"sa_opchar_admissable"} containing the
following elements
\itemize{
\item A tibble in the slot \code{$opchar} summarising the operating
characteristics of the supplied designs.
\item Each of the input variables as specified, subject to internal
modification.
}
}
\description{
\code{opchar_admissable()} supports the evaluation of the operating
characteristics of admissable group sequential single-arm clinical trial
designs for a single binary primary endpoint, determined using
\code{des_admissable()}.
}
\details{
For each value of \ifelse{html}{\out{<i>pi</i>}}{\eqn{\pi}} in
the supplied vector \ifelse{html}{\out{<b><i>pi</i></b>}}{\eqn{\bold{\pi}}},
\code{opchar_admissable()} evaluates the power, ESS, and other key
characteristics, of each of the supplied designs.
Calculations are performed conditional on the trial stopping in one of the
stages specified using the input (vector) \code{k}.
}
\examples{
# Find the admissable two-stage design for the default parameters
des <- des_admissable()
# Determine operating characteristics for a range of response probabilities
opchar <- opchar_admissable(des, pi = seq(0, 1, 0.01))
}
\seealso{
\code{\link{des_admissable}}, and their associated \code{plot}
family of functions.
}
|
cc14e1cf8319f8a967c45ffb049e2f54c126321b
|
4c671b09b63895596debc68f8a1eee4de7507356
|
/1_Names_Load Data Functions.R
|
17f53331e930c06952424ca882fb3f2133a35ce9
|
[] |
no_license
|
ryantimpe/Kerasaurs
|
80782e631f4a5df29895348bc840da9c9ebdfdfa
|
c289e61ce2d2da7bdee437ffd7cd622681e581c5
|
refs/heads/master
| 2021-05-03T23:22:23.042740
| 2018-12-13T21:51:51
| 2018-12-13T21:51:51
| 120,399,442
| 8
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,703
|
r
|
1_Names_Load Data Functions.R
|
#this file loads the packages and creates the functions that will be used in the model
require(readr)
require(stringr)
require(dplyr)
require(purrr)
require(tokenizers)
require(keras)
# This function loads the raw text of the _saurs
get_saurs <- function() {
readRDS("data/list_of_extinct_reptiles.RDS") %>%
str_replace_all("[^[:alnum:] ]", "") %>% # remove any special characters
tolower
}
add_stop <- function(saurs, symbol="+") {
str_c(saurs, symbol)
} # make a note for the end of a _saur name
# We want to predict each of the n character on the _saur name. Have to split one
# data point into n data points (where n is the number of characters on the plate).
# So _saur ABC would become data points "A", "AB", and "ABC"
split_into_subs <- function(saurs){
saurs %>%
tokenize_characters(lowercase=FALSE) %>%
map(function(saur) map(1:length(saur),function(i) saur[1:i])) %>%
flatten()
}
# make each data point the same number of characters by
# adding a padding symbol * to the font
fill_data <- function(saur_characters, max_length = 20){
saur_characters %>%
map(function(s){
if (max_length+1 > length(s)) {
fill <- rep("*", max_length+1 - length(s))
c(fill, s)
} else {
s
}
})
}
# convert the data into vectors that can be used by keras
vectorize <- function(data,characters, max_length){
x <- array(0, dim = c(length(data), max_length, length(characters)))
y <- array(0, dim = c(length(data), length(characters)))
for(i in 1:length(data)){
for(j in 1:(max_length)){
x[i,j,which(characters==data[[i]][j])] <- 1
}
y[i,which(characters==data[[i]][max_length+1])] <- 1
}
list(y=y,x=x)
}
|
83bbbbd00c095f9d751b7f0e90199a1583281456
|
5e968d1fa5ae2f0b3f15693561d61571a0b27a77
|
/R Code/Producemobile Time Series Analysis/producemobile_tsa.r
|
c9087445dccea46c37fb7e7718ba6db32ad373e7
|
[] |
no_license
|
adityagi1/everybody-eats-summer-2020
|
a4dadce45728122f046c926bbdfcbb96b85f2e5c
|
fbd17660736eae447a447bbb594bd1dea19e8da9
|
refs/heads/master
| 2022-12-28T22:06:29.840546
| 2020-10-20T02:50:39
| 2020-10-20T02:50:39
| 305,567,747
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 4,331
|
r
|
producemobile_tsa.r
|
library(tidyverse)
library(readxl)
library(lubridate)
library(imputeTS)
pm = read_excel("../../Data/ProduceMobile/Producemobile_historical_data_clean.xlsx",
sheet = "Sheet1")
#rename column names
names(pm) = c("date", "in_out", "amt_food_received","num_guests_signed_in",
"num_individuals_served", "truck_arrival_time","num_volunteers",
"leftover_pickup","num_boxes_pickup", "num_boxes_waste","note")
#form food received time series (has missing values)
amt_food_rec.ts = ts(pm$amt_food_received, frequency = 12)
plot(amt_food_rec.ts)
#form num individuals served time series (has missing values)
num_ind.ts = ts(pm$num_individuals_served, frequency = 12)
plot(num_ind.ts)
#form the num volunteers time series (has missing values)
num_volunteers.ts = ts(pm$num_volunteers, frequency = 12)
plot(num_volunteers.ts)
#num_ind, amt_food_rec, num_volunteers have na's that will need to be imputed
num_ind.ts.imp = na_seadec(num_ind.ts, algorithm = "interpolation")
plot(num_ind.ts.imp)
ggplot_na_imputations(num_ind.ts, num_ind.ts.imp)
#amount food received
amt_food_rec.ts.imp = na_seadec(amt_food_rec.ts, algorithm = "interpolation")
plot(amt_food_rec.ts.imp)
ggplot_na_imputations(amt_food_rec.ts, amt_food_rec.ts.imp)
#num_volunteers
num_volunteers.ts.imp = na_seadec(num_volunteers.ts, algorithm = "interpolation")
ggplot_na_imputations(num_volunteers.ts, num_volunteers.ts.imp)
#####Autocorrelation plots to check seasonality
#use smoothed data (using ma) as a preliminary estimate of the trend
num_ind.ts.ma = ma(num_ind.ts.imp, order = 12, centre = TRUE)
amt_food.ts.ma = ma(amt_food_rec.ts.imp, order = 12, centre = TRUE)
num_volunteers.ts.ma = ma(num_volunteers.ts.imp, order = 12, centre = TRUE)
#now compute the autocorr function of the de-trended data
acf(na_remove(num_ind.ts.imp - num_ind.ts.ma))
acf(na_remove(amt_food_rec.ts.imp - amt_food.ts.ma))
acf(na_remove(num_volunteers.ts.imp - num_volunteers.ts.ma))
#there are strong auto-correlations (nearing 0.4-0.5),
#indicating seasonality is present
######Time Series Decomposition
num_ind.decomp = decompose(num_ind.ts.imp, type = "additive")
num_volunteers.decomp = decompose(num_volunteers.ts.imp, type = "additive")
amt_food_rec.decomp = decompose(amt_food_rec.ts.imp, type = "additive")
#visualize decompositions
plot(num_ind.decomp)#, main = "Num. Individuals Decomposition")
plot(num_volunteers.decomp) #main = "Num. Volunteers Attending Decomposition")
plot(amt_food_rec.decomp)# main = "Amount food received Decomposition")
#there are strong seasonal and trend components, therefore it makes sense
#to use a holt-winters model
######MODELLING
num_ind.hw.mod = HoltWinters(num_ind.ts.imp, seasonal = "additive")
amt_food_rec.hw.mod = HoltWinters(amt_food_rec.ts.imp, seasonal = "additive")
num_volunteers.hw.mod = HoltWinters(num_volunteers.ts.imp, seasonal = "additive")
# let's look at summaries for the model
num_ind.hw.mod
amt_food_rec.hw.mod
num_volunteers.hw.mod
#####INTERPRETATIONS
#a very strong trend detected because beta = 0 for all three time series
##interpretations for seasonality
###NUM_IND_SERVED
#fewer individuals are served during:
#- Jan
#- Feb
#- April
#- october
#- december
#more individuals are served during:
#- march
#- may
#- september
#- november
### AMT_FOOD_RECEIVED
#lesser food is received during
#- february
#- april
#- june
#- october
#- december
#more food is received during:
#january,
#march,
#may,
#july,
#september,
#november
#more volunteers tend to come in:
#january
#february
#march
#april
#september
#november
#less volunteers tend to come in:
#may
#june
#july
#august
#october
#december
######PREDICTIONS
num_ind.hw.pred = predict(num_ind.hw.mod, n.ahead = 6, prediction.interval = TRUE)
amt_food_rec.hw.pred = predict(amt_food_rec.hw.mod, n.ahead = 6, prediction.interval = TRUE)
num_volunteers.hw.pred = predict(num_volunteers.hw.mod, n.ahead = 6, prediction.interval = TRUE)
##plot predictions
plot(num_ind.hw.mod, num_ind.hw.pred, type = 'b', main = "Forecasted Num. Individuals Arriving Till December 2020")
plot(amt_food_rec.hw.mod, amt_food_rec.hw.pred, type = 'b', main = "Forecasted Amt. Food Arriving Till December 2020")
plot(num_volunteers.hw.mod, num_volunteers.hw.pred, type = 'b', main = "Forecasted Num. Volunteers Till December 2020")
|
c01b07e4222a9a434cffcda6fe0dab792bd5099a
|
7188afce97d4674ec52fe33aeaf27b38e0889ac0
|
/plot2.R
|
fd59ad7864091381ec92e4c4f10f2163e9ef9280
|
[] |
no_license
|
joski/ExData_Plotting1
|
b7aaa92f0faa94bd343e0e7bf239595a9cd49f9d
|
a0392e468ac403d04d8727b08f57dacdb5b1bedd
|
refs/heads/master
| 2020-03-07T05:28:33.147234
| 2018-03-29T18:24:57
| 2018-03-29T18:24:57
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 243
|
r
|
plot2.R
|
source("load_data.R")
data<-load_data()
png("plot2.png", width=400, height=400)
with(data,plot(Time,Global_active_power,
type="l",
xlab = "",
ylab = "Global Active Power (kilowatts)"))
dev.off()
|
fae3732706ea29acd8534118f8ff641aba989bb2
|
d175a3870bbe9086724b614aaf681eda6002fa83
|
/server.R
|
a2a9e45db6ee1a729cdb7417ed1666669ff4e60c
|
[] |
no_license
|
MFKiani/ShinyAppCode
|
22bec721bdf85a026a050e4388627de5ade785ef
|
652981e69fc5cb35dffec5c555e96620c58c1c5e
|
refs/heads/master
| 2021-01-10T17:00:19.093394
| 2015-05-23T18:31:37
| 2015-05-23T18:31:37
| 36,120,606
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,229
|
r
|
server.R
|
library(shiny)
# We tweak the "am" field to have nicer factor labels. Since this doesn't
# rely on any user inputs we can do this once at startup and then use the
# value throughout the lifetime of the application
# Define server logic required to plot various variables against mpg
shinyServer(function(input, output) {
# Compute the forumla text in a reactive expression since it is
# shared by the output$caption and output$mpgPlot expressions
formulaText <- reactive({
paste("Normal distributions with mean", input$mean1, "and standard deviation", input$sd1,
"(in red) and mean",input$mean2, "and standard deviation", input$sd2, "(in blue)")
})
# Return the formula text for printing as a caption
output$caption1 <- renderText({
formulaText()
})
# Generate a plot of the requested variable against mpg and only
# include outliers if requested
output$normalPlot1 <- renderPlot({
x=seq(-10,10,length=200)
y1=dnorm(x,mean=input$mean1,sd=input$sd1)
plot(x,y1,type="l",lwd=2,col="red")
par(new = TRUE)
y2=dnorm(x,mean=input$mean2,sd=input$sd2)
plot(x,y2,type="l",lwd=2,col="blue")
})
})
|
6622733770139bc3c00ca3772d78cc176b5e736b
|
985015b0c5399a82a1c874e1d68406aeb8042c72
|
/dockerfiles/haloplex-qc/CoveragePlots.R
|
006e36bb92d74a3ed93949795cb5af36c90c5a51
|
[] |
no_license
|
genome/cle-myeloseq
|
f27f48c8e2fc868ac6a0ee3178c0f180be0871dc
|
fedd71f8daed318f3e4794e398797d038c002350
|
refs/heads/master
| 2021-10-10T00:53:51.361753
| 2021-10-08T16:34:19
| 2021-10-08T16:34:19
| 180,651,790
| 1
| 2
| null | 2021-05-04T19:04:01
| 2019-04-10T19:40:08
|
Perl
|
UTF-8
|
R
| false
| false
| 2,966
|
r
|
CoveragePlots.R
|
require(scales)
sample.name <- commandArgs(T)[1]
pdf(height=8.5,width=11,file=paste0(sample.name,".coverage_qc.pdf"))
layout(matrix(c(1,1,2:5),nrow=3,byrow = T),heights = c(.2,1,1))
op <- par(mar=c(2,2,2,2))
plot.new()
text(0,.5,labels = sample.name,cex=2,font=2,xpd=T,pos=4)
par(op)
cov <- read.table(paste0(sample.name,".coverage.txt"),stringsAsFactors = F)
cov$V5 <- gsub("_.+","",cov$V5)
genes <- levels(factor(cov$V5))
gene.chunks <- split(genes, ceiling(seq_along(genes)/10))
invisible(lapply(gene.chunks,
function(g){
x <- subset(cov,cov$V5 %in% g)
stats <- cbind(aggregate(x$V4 ~ x$V5,FUN = length),
aggregate(x$V4 ~ x$V5,FUN = mean),
aggregate(x$V4 ~ x$V5,FUN = median),
aggregate(x$V4 ~ x$V5,FUN = function(X){ length(which(X<50)) }))[,c(1,2,4,6,8)]
rownames(stats) <- stats[,1]
apply(stats[,1:5],1,function(X){ cat(c("COVERAGE",X,"\n"),sep="\t"); })
op <- par(mar=c(4.5,4,4.5,2))
boxplot(log2(x$V4+1) ~ x$V5,las=2,axes=F,col="light gray")
axis(1,at=1:nrow(stats),labels=rownames(stats),las=2)
yAx <- c(0,1,20,50,100,500,seq(1000,max(x$V4,1000),by=1000))
axis(2,at=log2(yAx+1),labels = yAx)
mtext("Coverage depth (by position)",side=2,line=2.5)
box()
abline(h=log2(c(20,50)+1),lty=2,lwd=3,col="red")
mtext(c("Mean",round(stats[,3],0)),at=c(-.25,1:nrow(stats)),line=2.75,cex=.8)
mtext(c("Median",round(stats[,4],0)),at=c(-.25,1:nrow(stats)),line=1.5,cex=.8)
mtext(c("<50x",round(stats[,5],0)),at=c(-.25,1:nrow(stats)),line=0.25,cex=.8)
invisible(par(op))
}))
invisible(dev.off())
invisible(require(scales))
pdf(height=11,width=8.5,file=paste0(sample.name,".gc_length_qc.pdf"))
layout(matrix(1:3,nrow=3,byrow = T),heights = c(.2,1,1))
op <- par(mar=c(2,2,2,2))
plot.new()
text(0,.5,labels = sample.name,cex=2,font=2,xpd=T,pos=4)
par(op)
x <- read.table(paste0(sample.name,".amplicon_counts.txt"),stringsAsFactors = F)
op <- par(mar=c(4,4,2,1),cex=1.25)
plot(x$V3-x$V2,log2(x$V5),pch=16,cex=.4,col=alpha("black",.5),xlab="Amplicon length",ylab="Read counts (log2)")
length.loess <- loess.smooth(x$V3-x$V2,log2(x$V5))
lines(length.loess,col="blue",lwd=3)
abline(h=log2(50),col="red",lwd=3,lty=3)
abline(h=log2(20),col="red",lwd=3,lty=3)
cat(paste0(c("LENGTHFITX\t",paste0(length.loess$x,collapse=","))))
cat("\n")
cat(paste0(c("LENGTHFITY\t",paste0(length.loess$y,collapse=","))))
cat("\n")
plot(x$V6*100,log2(x$V5),pch=16,cex=.4,col=alpha("black",.5),xlab="Amplicon GC%",ylab="Read counts (log2)")
gc.loess <- loess.smooth(x$V6*100,log2(x$V5))
lines(gc.loess,col="blue",lwd=3)
abline(h=log2(50),col="red",lwd=3,lty=3)
abline(h=log2(20),col="red",lwd=3,lty=3)
cat(paste0(c("GCFITX\t",paste0(gc.loess$x,collapse=","))))
cat("\n")
cat(paste0(c("GCFITY\t",paste0(gc.loess$y,collapse=","))))
cat("\n")
par(op)
invisible(dev.off())
|
1eadbb12f1feb597ecc4417e001eef1e1fb63f42
|
ffdea92d4315e4363dd4ae673a1a6adf82a761b5
|
/data/genthat_extracted_code/hawkes/examples/simulateHawkes.Rd.R
|
eda387c2ed25579725f6550cfc8ef0fa8e42b585
|
[] |
no_license
|
surayaaramli/typeRrh
|
d257ac8905c49123f4ccd4e377ee3dfc84d1636c
|
66e6996f31961bc8b9aafe1a6a6098327b66bf71
|
refs/heads/master
| 2023-05-05T04:05:31.617869
| 2019-04-25T22:10:06
| 2019-04-25T22:10:06
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 457
|
r
|
simulateHawkes.Rd.R
|
library(hawkes)
### Name: simulateHawkes
### Title: Hawkes process simulation Function
### Aliases: simulateHawkes
### ** Examples
#One dimensional Hawkes process
lambda0<-0.2
alpha<-0.5
beta<-0.7
horizon<-3600#one hour
h<-simulateHawkes(lambda0,alpha,beta,horizon)
#Multivariate Hawkes process
lambda0<-c(0.2,0.2)
alpha<-matrix(c(0.5,0,0,0.5),byrow=TRUE,nrow=2)
beta<-c(0.7,0.7)
horizon<-3600#one hour
h<-simulateHawkes(lambda0,alpha,beta,horizon)
|
41fe9f7b6f7180e00a09a589546b81344b98e8c6
|
20c506f33d3bfe2322d9be7a894d64bd6a179fff
|
/R code
|
3a8e10a2c2c3e308682852e15fbf60fd4a6ded9a
|
[] |
no_license
|
johnukfr/IDaSRP
|
ccdb75eab5bd7a42b63b2fd6a7f998d68599bbce
|
329bbd0260bcc62dacf5614ee7abcd642c189c9e
|
refs/heads/master
| 2020-07-04T18:04:08.063693
| 2020-05-15T13:57:21
| 2020-05-15T13:57:21
| 202,365,845
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 201,845
|
R code
|
#!/usr/bin/Rscript
# sink("IDaSRP_077.3.txt", type = c("output", "message"))
#####---------------------------------------------------------------------------
##### R code for Disertation 'Information Demand and Stock Return Predictability'
##### University of Essex
##### Written by: Jonathan Legrand
#####---------------------------------------------------------------------------
print("R code Version 77 for Disertation 'Information Demand and Stock Return Predictability': University of Essex: Written by: Jonathan Legrand. Written to be ran on Univeristy ceres.essex.ac.uk")
####---------------------------------------------------------------------------
#### Code preperation (clear memory, prepare libraries)
####---------------------------------------------------------------------------
print("Code preperation (clear memory, prepare libraries)")
# remove all objects in the memory
rm(list = ls())
# # Install pakages/libraries/dependancies needed
# install.packages(c("readxl", "xts", "fpp", "astsa", "tidyverse", "dplyr", "rugarch", "Metrics", "e1071", "forecast", "R.utils", "aTSA", "R.utils", "EnvStats", "ggplot2", "plotly", "tsoutliers")
# Import Libraries
library(readxl)
# The library 'astsa' is used for Econometric Modelling. See more at:
#https://www.rdocumentation.org/packages/astsa/versions/1.6/
library(astsa)
# The library 'rugarch' is used for GARCH Modelling. See more at:
# https://stats.stackexchange.com/questions/93815/fit-a-garch-1-1-model-with-covariates-in-r
# https://cran.r-project.org/web/packages/rugarch/rugarch.pdf
# https://rdrr.io/rforge/rugarch/man/ugarchspec-methods.html
library(rugarch)
library(tidyverse)
library(dplyr)
library(xts)
# The library 'Metrics' allows for the calculation of RMSE
library(Metrics)
# The library 'e1071' allows for the calculation of skewness
library(e1071)
# The library 'aTSA' allows for the calculation of the ADF test
library(aTSA)
library(forecast)
# R.utils is needed for timeout functions
library(R.utils)
# EnvStats is needed to plot pdf's
library(EnvStats)
# plotly will allow us to plot 3d graphs
library(plotly)
Sys.setenv("plotly_username"="johnukfr")
# Sys.setenv("plotly_api_key"=)
library(ggplot2)
theme_set(theme_minimal())
# tsoutliers allows us to perform jarque-bera tests
library(tsoutliers)
####---------------------------------------------------------------------------
#### Set Datasets
####---------------------------------------------------------------------------
print('Set Datasets')
###---------------------------------------------------------------------------
### SPX (i.e.: the S&P 500)
###---------------------------------------------------------------------------
SPX_Dates_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/1-Month_Treasury_Rate/1-Month_Treasury_Constant_Maturity_Rate_FRED_id_DGS1MO_2004.01.01_to_2019.03.13.xlsx",
sheet = "Without_FRED_or_O-M_Holidays",
range = "A2:A3794",
col_names = "SPX_Dates")
SPX_Dates = as.Date(SPX_Dates_df$SPX_Dates,"%Y-%m-%d", tz="Europe/London")
# Setup Dates for subsets:
SPX_Dates_df_m1 = slice(SPX_Dates_df, 2:3793)
SPX_Datesm1 = as.Date(SPX_Dates_df_m1$SPX_Dates,"%Y-%m-%d", tz="Europe/London")
SPX_Dates_df_m2 = slice(SPX_Dates_df, 3:3793)
SPX_Datesm2 = as.Date(SPX_Dates_df_m2$SPX_Dates,"%Y-%m-%d", tz="Europe/London")
# For a list of available time zones see:
# https://en.wikipedia.org/wiki/List_of_tz_database_time_zones
# # One can see the list of dates with the comand
# fix(SPX_Dates)
DGS1MO_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/1-Month_Treasury_Rate/1-Month_Treasury_Constant_Maturity_Rate_FRED_id_DGS1MO_2004.01.01_to_2019.03.13.xlsx",
sheet = "Without_FRED_or_O-M_Holidays",
range = "B2:B3794",
col_names = "DGS1MO")
DGS1MO_df_dates = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/1-Month_Treasury_Rate/1-Month_Treasury_Constant_Maturity_Rate_FRED_id_DGS1MO_2004.01.01_to_2019.03.13.xlsx",
sheet = "Without_FRED_or_O-M_Holidays",
range = "A2:A3794",
col_names = "DGS1MO_Dates")
DGS1MO_df_dates = as.Date(DGS1MO_df_dates$DGS1MO_Dates,"%Y-%m-%d", tz="Europe/London")
DGS1MO = as.matrix(DGS1MO_df)
DTB3_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/3-Month_Treasury_Bill/Daily_3-Month_Treasury_Bill_FRED_id_TB3MS_2004.01.01_to_2019.03.13.xlsx",
sheet = "Without_FRED_or_O-M_Holidays",
range = "B2:B3794",
col_names = "TB3MS")
DTB3 = as.matrix(DTB3_df)
DDGS10_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/10-Year_Treasury_Rate/10-Year_Treasury_Constant_Maturity_Rate_FRED_id_DGS10_2004.01.01_to_2019.03.13.xlsx",
sheet = "Without_FRED_or_O-M_Holidays",
range = "B2:B3794",
col_names = "DGS10")
DDGS10 = as.matrix(DDGS10_df)
SPX_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/Realised Volatility/5min_realised_volatility_from_Oxford-Man.xlsx",
sheet = "SPX2004-2019.03.1WithoutFREDHol",
range = 'T2:T3794',
col_names = "SPX")
SPX = as.matrix(SPX_df)
SPX_RVar_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/Realised Volatility/5min_realised_volatility_from_Oxford-Man.xlsx",
sheet = "SPX2004-2019.03.1WithoutFREDHol",
range = "I2:I3794",
col_names = "SPX_RVar")
# Note that the Ox-Man institute provides real VARIANCE as per
# https://realized.oxford-man.ox.ac.uk/documentation/estimators
SPX_RV_df = (SPX_RVar_df^0.5)
SPX_RV = as.matrix(SPX_RV_df)
colnames(SPX_RV) = c('SPX_RV')
SPX_RV_zoo=zoo(SPX_RV, as.Date(SPX_Dates))
###---------------------------------------------------------------------------
### FTSE
###---------------------------------------------------------------------------
print("Importing FTSE data")
FTSE_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/Realised Volatility/5min_realised_volatility_from_Oxford-Man.xlsx",
sheet = "FTSE2001-2019.06.21",
range = 'U2:U4910',
col_names = "FTSE")
FTSE_matrix = as.matrix(FTSE_df)
FTSE_Dates = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/Realised Volatility/5min_realised_volatility_from_Oxford-Man.xlsx",
sheet = "FTSE2001-2019.06.21",
range = 'A2:A4910',
col_names = "FTSE Dates")
FTSE_Dates_matrix = as.matrix(FTSE_Dates)
FTSE_zoo = as.zoo(FTSE_matrix, as.Date(FTSE_Dates_matrix))
# # Band of England (BoE) r_f (risk free bond daily rate of return)
# Zero coupon nominal curves as defined by the BoE: : The spot interest rate or zero coupon yield is the rate at which an individual cash flow on some future date is discounted to determine its present value. By definition it is the yield to maturity of a zero coupon bond and can be considered as an average of single period rates to that maturity. Conventional dated stocks with a significant amount in issue and having more than three months to maturity, plus General Collateral repo rates (at the short end) are used to estimate these yields; index-linked stocks, irredeemable stocks, double dated stocks, stocks with embedded options, variable and floating stocks are all excluded.
# Data gathered from http://www.bankofengland.co.uk/boeapps/iadb/index.asp?Travel=NIxIRx&levels=1&XNotes=Y&C=2C6&C=RN&C=DR6&G0Xtop.x=57&G0Xtop.y=9&XNotes2=Y&Nodes=X4051X4052X4053X4054X4066X4067X4068X38263&SectionRequired=I&HideNums=-1&ExtraInfo=true#BM
# https://www.bankofengland.co.uk/statistics/details/further-details-about-yields-data
# Note that: The prices of conventional gilts are quoted in terms of £100 nominal. As per: https://www.dmo.gov.uk/responsibilities/gilt-market/about-gilts/
print("Importing BoE data")
BoE5YR_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/BoE_5_Year_Gilt/British_ Government_Securities_Yield_5_year_Nominal_Zero_Coupon.xlsx",
sheet = "British_ Government_Securities_",
range = 'A5:B3937',
col_names = c("BoE 5Y Yield Dates", "BoE 5Y Yield"))
BoE5YR_matrix = as.matrix(BoE5YR_df)
BoE5YR_zoo = zoo(BoE5YR_matrix[1:3933,2], as.Date(BoE5YR_matrix[1:3933,1]))
SVIFTSE_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/SVI/FTSE 100 World/SVI_from_2004.01.01_to_2019.03.13_normalised_by_2016.06.24_all_days.xlsx",
sheet = "Sheet1",
range = "A2:B5552",
col_names = c("SVIFTSE_dates", "SVIFTSE"))
SVIFTSE_zoo = zoo(SVIFTSE_df[1:5551,2], as.Date(as.matrix(SVIFTSE_df[1:5551,1])))
SVIFTSE = as.matrix(SVIFTSE_zoo)
RVarFTSE_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/Realised Volatility/5min_realised_volatility_from_Oxford-Man.xlsx",
sheet = "FTSE2001-2019.06.21",
range = 'J2:J4910',
col_names = "FTSE")
# Note that the Ox-Man institute provides realised VARIANCE as per
# https://realized.oxford-man.ox.ac.uk/documentation/estimators
RVFTSE_df = (RVarFTSE_df^0.5)
RVFTSE_zoo = zoo(RVFTSE_df, as.Date(FTSE_Dates_matrix))
####---------------------------------------------------------------------------
#### Derive variables
####---------------------------------------------------------------------------
print("Derive variables")
###---------------------------------------------------------------------------
### SPX
###---------------------------------------------------------------------------
# Several points have to be made about the FRED's data at this point:
#
# 1st:
# The 1-, 2-, and 3-month rates are equivalent to the 30-, 60-,
# and 90-day dates respectively, reported on the Board's Commercial Paper Web
# page (www.federalreserve.gov/releases/cp/). This is as per the FRED's own
# website (https://fred.stlouisfed.org/series/DGS1MO#0)'s referance
# (https://www.federalreserve.gov/releases/h15/current/h15.pdf).
# Figures are annualized using a 360-day year or bank interest as per
# https://www.federalreserve.gov/releases/h15/.
# We are using FRED's Constant Maturity Rate (CMR) data, more info on that:
# https://www.investopedia.com/terms/c/constantmaturity.asp
# https://www.investopedia.com/terms/c/cmtindex.asp
# https://fred.stlouisfed.org/release/tables?rid=18&eid=289&snid=316
#
# 2nd: From there,
# we workout bellow the unit return from holding a one-month Treasury bill over
# the period from t-1 to t by calculating its differrance in daily price
# where: Daily price = (((CMR/100)+1)^(-(1/12)))*1000
#
US_1MO_r_f=((((((DGS1MO/100)+1)^(-1/12))*1000)-
((((lag(DGS1MO)/100)+1)^(-1/12))*1000))/
((((lag(DGS1MO)/100)+1)^(-(1/12)))*1000))
US_1MO_r_f_zoo = zoo(US_1MO_r_f, as.Date(DGS1MO_df_dates))
# Construct the returns variable, R_t (and its laged value, R_tm1)
# Note things here:
#
# 1st: that due to the diferancing nessesary to calculate 'R',
# the first value is empty. The comand:
# as.matrix(((INDEX_df-lag.xts(data.matrix(INDEX_df)))/INDEX_df)-APPROPRIATE_r_f)
# Where INDEX is SPX or FTSE and APPROPRIATE_r_f is
# US_1MO_r_f or BoE_5Y_r_f displays this well.
#
# 2nd: In order for the correct 'R' values to be associated with the correct
# dates, the element 'R' has to be changed into a 'zoo' element.
# This will allow future models to have data points with correct date lables.
# A vecor element is left however,
# for libraries and functions that do not support them.
#
SPX_R_matrix = as.matrix(((SPX_df-lag.xts(data.matrix(SPX_df)))
/lag.xts(data.matrix(SPX_df)))
-US_1MO_r_f)[2:length(SPX)]
SPX_R_zoo = zoo(SPX_R_matrix,as.Date(SPX_Datesm1))
##---------------------------------------------------------------------------
## SVI1
##---------------------------------------------------------------------------
# The SVI Excel file tends to have empty values at its end.
# The 'slice' function bellow will remove them.
SPX_SVI_df = slice(read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/SVI/SVI_from_2004.01.01_to_2019.03.13_normalised_by_2018.02.06_Only_Trading_Days.xlsx",
sheet = "SVI_Without_O-M_Holidays",
range = "A1:B3802",
col_names = c("Dates", "SPX_SVI")),
1:length(SPX_Dates))
SPX_SVI = as.matrix(SPX_SVI_df$SPX_SVI) # Convert the data frame 'SPX_SVI_df' into a matrix
SPX_SVI_zoo = zoo(SPX_SVI, as.Date(SPX_SVI_df$Dates,"%Y-%m-%d", tz="Europe/London"))
# Construct Google's Search Vecrot Index (SVI).
SPX_dSVI = as.matrix(SPX_SVI - lag.xts(data.matrix(SPX_SVI)))[2:length(SPX_SVI)]
colnames(SPX_dSVI) = c('SPX_dSVI')
SPX_dSVI_zoo = zoo(SPX_dSVI, as.Date(Dates[1:3792]))
##---------------------------------------------------------------------------
## SVI2
##---------------------------------------------------------------------------
SPX_SVI2_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/SVI/2/SVI_from_2004.01.01_to_2019.03.13_normalised_by_2018.02.06_all_days.xlsx",
sheet = "Sheet1",
range = "A2:B5552",
col_names = c("SVI2_dates", "SPX_SVI2"))
SPX_SVI2_zoo = zoo(SPX_SVI2_df[1:5551,2], as.Date(as.matrix(SPX_SVI2_df[1:5551,1])))
SPX_SVI2 = as.matrix(SPX_SVI2_zoo)
SPX_dSVI2_zoo = zoo(SPX_SVI2 - lag.xts(data.matrix(SPX_SVI2)),
as.Date(as.matrix(SPX_SVI2_df[1:5551,1])))
SPX_dSVI2_zoo = SPX_dSVI2_zoo - (SPX_R_zoo*0)
colnames(SPX_dSVI2_zoo) = c('SPX_dSVI2')
##---------------------------------------------------------------------------
## SVI CVP
##---------------------------------------------------------------------------
SPX_SVICPV_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/SVI/CPV/SVI_from_2004.01.01_to_2017.03.15_normalised_by_2018.02.06 _CPV_days.xlsx",
sheet = "Sheet1",
range = "A2:B4824",
col_names = c("SPX_SVICPV_dates", "SPX_SVICPV"))
SPX_SVICPV_zoo = zoo(SPX_SVICPV_df[1:4823,2],
as.Date(as.matrix(SPX_SVICPV_df[1:4823,1])))
SPX_SVICPV = as.matrix(SPX_SVICPV_zoo)
SPX_dSVICPV_all_days_zoo=zoo(SPX_SVICPV-lag.xts(data.matrix(SPX_SVICPV)),
as.Date(as.matrix(SPX_SVICPV_df[1:4823,1])))
SPX_dSVICPV = SPX_dSVICPV_all_days_zoo - (SPX_R_zoo*0)
colnames(SPX_dSVICPV) = c('SPX_dSVICPV')
##---------------------------------------------------------------------------
## SVI CVP US
##---------------------------------------------------------------------------
SPX_SVICPVUS_df = read_excel("C:/Users/johnukfr/OneDrive/UoE/Disertation/Data/SVI/CPV_US/SVI_from_2004.01.01_to_2017.03.15_normalised_by_2016.11.09_all_days.xlsx",
sheet = "Sheet1",
range = "A2:B4824",
col_names = c("SPX_SVICPV_dates", "SPX_SVICPV"))
SPX_SVICPVUS_zoo = zoo(SPX_SVICPVUS_df[1:4823,2],
as.Date(as.matrix(SPX_SVICPVUS_df[1:4823,1])))
SPX_SVICPVUS_matrix = as.matrix(SPX_SVICPVUS_zoo)
SPX_dSVICPVUS_all_days_zoo=zoo(SPX_SVICPVUS_matrix-lag.xts(SPX_SVICPVUS_matrix),
as.Date(as.matrix(SPX_SVICPVUS_df[1:4823,1])))
SPX_dSVICPVUS_R = SPX_dSVICPVUS_all_days_zoo-(SPX_R_zoo*0)
colnames(SPX_dSVICPVUS_R) = c('SPX_dSVICPVUS_R')
SPX_dSVICPVUS_Rm1 = lag(SPX_dSVICPVUS_all_days_zoo) - (SPX_R_zoo*0)
colnames(SPX_dSVICPVUS_R) = c('SPX_dSVICPVUS_Rm1')
#---------------------------------------------------------------------------
# # SVI CPV MA
#---------------------------------------------------------------------------
# the dSVI data is quite noisy, so we use a Moving Average to smooth it out
# SPX_MAdSVICPV_all_days_zoo=zoo(movavg(SPX_dSVICPV_all_days_zoo, M,type="s"),
# as.Date(as.matrix(SPX_SVICPV_df[1:4823,1])))
# SPX_MAdSVICPV_all_days_zoo=zoo(SPX_dSVICPV_all_days_zoo, as.Date(as.matrix(SPX_SVICPV_df[1:4823,1])))
# SPX_MAdSVICPV = na.omit(SPX_MAdSVICPV_all_days_zoo) - (SPX_R_zoo*0)
# Deppending on the M chosen, the sample sie differs since we loose M starting
# datapoints; so to insure we get comparable data throughout, we will only start
# R and MAdSVI on 2004-01-07:
# SPX_R_Sample_zoo = SPX_R_zoo[3:length(SPX_R_zoo)]
# SPX_R_Sample_zoo = SPX_R_Sample_zoo - (SPX_MAdSVICPV*0)
# SPX_MAdSVICPV = SPX_MAdSVICPV - (SPX_R_Sample_zoo*0)
# SPX_dSVICPV = SPX_MAdSVICPV - (SPX_R_Sample_zoo*0)
# SPX_Sample_Dates = as.Date(zoo::index(SPX_MAdSVICPV))
# Note that when usinf MAdSVI, the sample would start M days after 2004-01-01;
# the maximum we will allow M to be will be 6, such that it starts on
# 2004-01-07 and ends on 2017-03-15.
# Its last in-GARCH-training-sample value is 2004-12-31 and is its 245th value.
# Its first out-of-GARCH-training-sample value is 2005-01-03 and is its 246th value.
###---------------------------------------------------------------------------
### FTSE (including SVIFTSE)
###---------------------------------------------------------------------------
BoE_r_f=((((((as.numeric(BoE5YR_zoo)/100)+1)^(-5))*100)-
((((lag(as.numeric(BoE5YR_zoo))/100)+1)^(-5))*100))/
((((lag(as.numeric(BoE5YR_zoo))/100)+1)^(-5)*100)))
BoE_r_f_zoo = zoo(BoE_r_f[2:length(BoE_r_f)],
as.Date(BoE5YR_matrix[2:length(BoE_r_f),1]))
BoE_p_f=((((as.numeric(BoE5YR_zoo)/100)+1)^(-5))*100)
cumulative_BoE_r_f = matrix(,nrow=length(BoE_r_f_zoo))
for (i in c(1:length(cumulative_BoE_r_f)))
{cumulative_BoE_r_f[i] = prod((1+BoE_r_f)[1:i], na.rm=TRUE)}
FTSE_R_zoo = zoo(na.exclude(as.matrix((FTSE_df-lag.xts(data.matrix(FTSE_df)))
/lag.xts(data.matrix(FTSE_df)))),
as.Date(FTSE_Dates_matrix[2:4909])) - BoE_r_f_zoo
FTSE_R_matrix = as.matrix(FTSE_R_zoo)
# # Construct Google's Search Vecrot Index (SVI).
dSVIFTSE=matrix(data.matrix(data.matrix(SVIFTSE)-
lag.xts(data.matrix(SVIFTSE)))[2:length(SVIFTSE)])
colnames(dSVIFTSE) = c('dSVIFTSE')
dSVIFTSE_zoo = zoo(dSVIFTSE,
as.Date(as.matrix(SVIFTSE_df[2:5551,1])))
dSVIFTSE_zoo = dSVIFTSE_zoo - (FTSE_R_zoo*0)
colnames(dSVIFTSE_zoo) = c('dSVIFTSE')
dSVIFTSE=matrix(dSVIFTSE_zoo)
Dates = as.Date(rownames(as.matrix(dSVIFTSE_zoo)))
####---------------------------------------------------------------------------
#### Descriptive Statistics of variables
####---------------------------------------------------------------------------
print("Descriptive Statistics of variables")
###---------------------------------------------------------------------------
### FTSE
###---------------------------------------------------------------------------
# FTSE excess return
print("plot of FTSE Excess Returns")
epdfPlot(as.numeric(FTSE_R_zoo), main = "", xlab="")
pdfPlot(param.list = list(mean=0, sd=sd(FTSE_R_matrix)), add = TRUE, pdf.col = "red")
pdfPlot(add = TRUE, pdf.col = "blue")
legend("topright", legend = c("EPDF plot of Excess FTSE Returns", "N[0,Var(Excess FTSE Returns)]", "N(0,1)"), col = c("black", "red", "blue"), lwd = 2 * par("cex"))
title("PDF Plots")
print("Mean of FTSE_R_zoo")
mean(FTSE_R_zoo)
print("Standard Deviation of FTSE_R_zoo")
sd(FTSE_R_zoo)
print("Median of FTSE_R_zoo")
median(FTSE_R_zoo)
print("Skewness of FTSE_R_zoo")
skewness(as.numeric(FTSE_R_zoo))
print("Kurtosis of FTSE_R_zoo")
kurtosis(as.numeric(FTSE_R_zoo))
print("AutoCorrelation Function of FTSE_R_zoo")
acf(matrix(FTSE_R_zoo), lag.max = 1, plot = FALSE)
# FTSE's SVI
print("Mean of SVIFTSE_zoo")
mean(SVIFTSE_zoo) # Note that that's the same as 'mean(SPX_dSVI_zoo)'; after much study on the matter, it seems as though this mean and other statistics are correct, which only highlights how different seperate SVI draws can be.
print("Standard Deviation of SVIFTSE_zoo")
sd(SVIFTSE_zoo)
print("Median of SVIFTSE_zoo")
median(SVIFTSE_zoo)
print("Skewness of SVIFTSE_zoo")
skewness(as.numeric(SVIFTSE_zoo))
print("Kurtosis of SVIFTSE_zoo")
kurtosis(as.numeric(SVIFTSE_zoo))
print("AutoCorrelation Function of SVIFTSE_zoo")
acf(as.numeric(SVIFTSE_zoo), lag.max = 1, plot = FALSE)
print("Augmented Dickey-Fuller Test of SVIFTSE_zoo")
adf.test(matrix(SVIFTSE_zoo), nlag = NULL, output = TRUE)
# Our ADF test here should show that we reject the Null Hypothesis (low p-value)
# in preference for the alternative of stationarity. One can see that dSVIFTSE_zoo
# is stationary by ploting it with:
plot(SVIFTSE_zoo, xlab='Date', ylab='SVICPVUSSPX', col='black', type='l',
# main='dSVIFTSE'
)
print("PP Test of SVIFTSE_zoo")
PP.test(SVIFTSE_zoo)
print("KPSS Test of SVIFTSE_zoo")
kpss.test(matrix(SVIFTSE_zoo))
# FTSE's dSVI
print("Mean of dSVIFTSE")
mean(dSVIFTSE_zoo)
print("Standard Deviation of dSVIFTSE")
sd(dSVIFTSE_zoo)
print("Median of dSVIFTSE")
median(dSVIFTSE_zoo)
print("Skewness of dSVIFTSE")
skewness(as.numeric(dSVIFTSE_zoo))
print("Kurtosis of dSVIFTSE")
kurtosis(as.numeric(dSVIFTSE_zoo))
print("AutoCorrelation Function of dSVIFTSE")
acf(matrix(dSVIFTSE_zoo), lag.max = 1, plot = FALSE)
print("Augmented Dickey-Fuller Test of dSVIFTSE")
adf.test(matrix(dSVIFTSE_zoo), nlag = NULL, output = TRUE)
# Our ADF test here should show that we reject the Null Hypothesis (low p-value)
# in preference for the alternative of stationarity. One can see that dSVIFTSE_zoo
# is stationary by ploting it with:
plot(dSVIFTSE_zoo, xlab='Date', ylab='dSVIFTSE', col='black',
main='dSVIFTSE', type='l')
print("PP Test of dSVIFTSE")
PP.test(dSVIFTSE_zoo)
print("KPSS Test of dSVIFTSE")
kpss.test(matrix(dSVIFTSE_zoo))
plot(SVIFTSE_zoo, xlab='Date', ylab='SVIFTSE', col='blue',
# main='Google Search Vector Index (SVI) for the FTSE index over time',
type='l')
# FTSE's Realised Volatility
print("Mean of RVFTSE")
mean(RVFTSE_zoo)
print("Standard Deviation of RVFTSE")
sd(RVFTSE_zoo)
print("Median of RVFTSE")
median(RVFTSE_zoo)
print("Skewness of RVFTSE")
skewness(as.numeric(RVFTSE_zoo))
print("Kurtosis of RVFTSE")
kurtosis(as.numeric(RVFTSE_zoo))
print("AutoCorrelation Function of RVFTSE")
acf(as.numeric(RVFTSE_zoo), lag.max = 1, plot = FALSE)
# FTSE's BoE_r_f
print("Mean of BoE_r_f")
mean(BoE_r_f_zoo)
print("Standard Deviation of BoE_r_f")
sd(BoE_r_f_zoo)
print("Median of BoE_r_f")
median(BoE_r_f_zoo)
print("Skewness of BoE_r_f")
skewness(as.numeric(BoE_r_f_zoo))
print("Kurtosis of BoE_r_f")
kurtosis(as.numeric(BoE_r_f_zoo))
print("AutoCorrelation Function of BoE_r_f")
acf(as.numeric(BoE_r_f_zoo), lag.max = 1, plot = FALSE)
# Descriptive Plot
plot(SVIFTSE_zoo, xlab='Date', ylab='SVIFTSE', col='blue',
main='Google Search Vector Index (SVI) for the FTSE index over time', type='l')
###---------------------------------------------------------------------------
### SPX
###---------------------------------------------------------------------------
# SPX excess return
print("plot of SPX Excess Returns")
epdfPlot(as.numeric(SPX_R_zoo), main = "", xlab="")
lines(density(FTSE_R_matrix),col="orange", lwd = 2 * par("cex"))
pdfPlot(param.list = list(mean=0, sd=sd(FTSE_R_matrix)), add = TRUE, pdf.col = "red")
pdfPlot(param.list = list(mean=0, sd=sd(SPX_R_zoo)), add = TRUE, pdf.col = "green")
pdfPlot(add = TRUE, pdf.col = "blue")
legend("topright", legend = c("EPDF plot of Excess SPX Returns", "EPDF plot of Excess FTSE Returns", "N[0,Var(Excess FTSE Returns)]", "N[0,Var(Excess SPX Returns)]", "N(0,1)"), col = c("black", "orange", "red","green", "blue"), lwd = 2 * par("cex"))
# title("PDF Plots")
print("Mean of SPX_R_zoo")
mean(SPX_R_zoo)
print("Standard Deviation of SPX_R_zoo")
sd(SPX_R_zoo)
print("Median of SPX_R_zoo")
median(SPX_R_zoo)
print("Skewness of SPX_R_zoo")
skewness(as.numeric(SPX_R_zoo))
print("Kurtosis of SPX_R_zoo")
kurtosis(as.numeric(SPX_R_zoo))
print("AutoCorrelation Function of SPX_R_zoo")
acf(matrix(SPX_R_zoo), lag.max = 1, plot = FALSE)
# SPX's Realised Volatility
print("Mean of SPX_RV")
mean(SPX_RV_zoo)
print("Standard Deviation of SPX_RV")
sd(SPX_RV_zoo)
print("Median of SPX_RV")
median(SPX_RV_zoo)
print("Skewness of SPX_RV")
skewness(as.numeric(SPX_RV_zoo))
print("Kurtosis of SPX_RV")
kurtosis(as.numeric(SPX_RV_zoo))
print("AutoCorrelation Function of SPX_RV")
acf(as.numeric(SPX_RV_zoo), lag.max = 1, plot = FALSE)
# SPX's r_f
print("Mean of US_1MO_r_f")
mean(US_1MO_r_f[2:length(US_1MO_r_f)])
print("Standard Deviation of US_1MO_r_f")
sd(US_1MO_r_f[2:length(US_1MO_r_f)])
print("Median of US_1MO_r_f")
median(US_1MO_r_f[2:length(US_1MO_r_f)])
print("Skewness of US_1MO_r_f")
skewness(as.numeric(US_1MO_r_f[2:length(US_1MO_r_f)]))
print("Kurtosis of US_1MO_r_f")
kurtosis(as.numeric(US_1MO_r_f[2:length(US_1MO_r_f)]))
print("AutoCorrelation Function of US_1MO_r_f")
acf(as.numeric(US_1MO_r_f[2:length(US_1MO_r_f)]), lag.max = 1, plot = FALSE)
# SPX's SVI1
print("Mean of SPX_SVI_zoo")
mean(SPX_SVI_zoo) # Note that that's the same as 'mean(SPX_dSVI_zoo)'; after much study on the matter, it seems as though this mean and other statistics are correct, which only highlights how different seperate SVI draws can be.
print("Standard Deviation of SPX_SVI_zoo")
sd(SPX_SVI_zoo)
print("Median of SPX_SVI_zoo")
median(SPX_SVI_zoo)
print("Skewness of SPX_SVI_zoo")
skewness(as.numeric(SPX_SVI_zoo))
print("Kurtosis of SPX_SVI_zoo")
kurtosis(as.numeric(SPX_SVI_zoo))
print("AutoCorrelation Function of SPX_SVI_zoo")
acf(as.numeric(SPX_SVI_zoo), lag.max = 1, plot = FALSE)
print("Augmented Dickey-Fuller Test of SPX_SVI_zoo")
adf.test(matrix(SPX_SVI_zoo), nlag = NULL, output = TRUE)
# Our ADF test here should show that we reject the Null Hypothesis (low p-value)
# in preference for the alternative of stationarity. One can see that dSVIFTSE_zoo
# is stationary by ploting it with:
plot(SPX_SVI_zoo, xlab='Date', ylab='SVI1SPX', col='black', type='l',
# main='dSVIFTSE'
)
print("PP Test of SPX_SVI_zoo")
PP.test(SPX_SVI_zoo)
print("KPSS Test of SPX_SVI_zoo_zoo")
kpss.test(matrix(SPX_SVI_zoo))
# SPX's dSVI1
print("Mean of SPX_dSVI")
mean(SPX_dSVI_zoo) # Note that that's the same as 'mean(SPX_dSVI_zoo)'; after much study on the matter, it seems as though this mean and other statistics are correct, which only highlights how different seperate SVI draws can be.
print("Standard Deviation of SPX_dSVI")
sd(SPX_dSVI)
print("Median of SPX_dSVI")
median(SPX_dSVI)
print("Skewness of SPX_dSVI")
skewness(as.numeric(SPX_dSVI))
print("Kurtosis of SPX_dSVI")
kurtosis(as.numeric(SPX_dSVI))
print("AutoCorrelation Function of SPX_dSVI")
acf(as.numeric(SPX_dSVI), lag.max = 1, plot = FALSE)
print("Augmented Dickey-Fuller Test of SPX_dSVI")
adf.test(matrix(SPX_dSVI), nlag = NULL, output = TRUE)
# Our ADF test here should show that we reject the Null Hypothesis (low p-value)
# in preference for the alternative of stationarity. One can see that dSVIFTSE_zoo
# is stationary by ploting it with:
plot(SPX_dSVI_zoo, xlab='Date', ylab='dSVI1SPX', col='black', type='l',
# main='dSVIFTSE'
)
print("PP Test of SPX_dSVI")
PP.test(SPX_dSVI)
print("KPSS Test of SPX_dSVI_zoo")
kpss.test(matrix(SPX_dSVI_zoo))
# SPX's SVI2
print("Mean of SPX_SVI2_zoo")
mean(SPX_SVI2_zoo) # Note that that's the same as 'mean(SPX_dSVI_zoo)'; after much study on the matter, it seems as though this mean and other statistics are correct, which only highlights how different seperate SVI draws can be.
print("Standard Deviation of SPX_SVI2_zoo")
sd(SPX_SVI2_zoo)
print("Median of SPX_SVI2_zoo")
median(SPX_SVI2_zoo)
print("Skewness of SPX_SVI2_zoo")
skewness(as.numeric(SPX_SVI2_zoo))
print("Kurtosis of SPX_SVI2_zoo")
kurtosis(as.numeric(SPX_SVI2_zoo))
print("AutoCorrelation Function of SPX_SVI2_zoo")
acf(as.numeric(SPX_SVI2_zoo), lag.max = 1, plot = FALSE)
print("Augmented Dickey-Fuller Test of SPX_SVI2_zoo")
adf.test(matrix(SPX_SVI2_zoo), nlag = NULL, output = TRUE)
# Our ADF test here should show that we reject the Null Hypothesis (low p-value)
# in preference for the alternative of stationarity. One can see that dSVIFTSE_zoo
# is stationary by ploting it with:
plot(SPX_SVI2_zoo, xlab='Date', ylab='SVI2SPX', col='black', type='l',
# main='dSVIFTSE'
)
print("PP Test of SPX_SVI2_zoo")
PP.test(SPX_SVI2_zoo)
print("KPSS Test of SPX_SVI2_zoo")
kpss.test(matrix(SPX_SVI2_zoo))
# SPX's SPX_dSVI2_zoo
print("Mean of SPX_dSVI2")
mean(SPX_dSVI2_zoo)
print("Standard Deviation of SPX_dSVI2")
sd(SPX_dSVI2_zoo)
print("Median of SPX_dSVI2")
median(SPX_dSVI2_zoo)
print("Skewness of SPX_dSVI2")
skewness(as.numeric(SPX_dSVI2_zoo))
print("Kurtosis of SPX_dSVI2")
kurtosis(as.numeric(SPX_dSVI2_zoo))
print("AutoCorrelation Function of SPX_dSVI2")
acf(as.numeric(SPX_dSVI2_zoo), lag.max = 1, plot = FALSE)
print("Augmented Dickey-Fuller Test of SPX_dSVI2_zoo")
adf.test(matrix(SPX_dSVI2_zoo), nlag = NULL, output = TRUE)
plot(SPX_dSVI2_zoo, xlab='Date', ylab='dSVI2SPX', col='black', type='l',
# main='dSVIFTSE'
)
print("PP Test of SPX_dSVI2_zoo")
PP.test(SPX_dSVI2_zoo)
print("KPSS Test of SPX_dSVI2_zoo")
kpss.test(matrix(SPX_dSVI2_zoo))
# SPX's SVICPV
print("Mean of SPX_SVICPV_zoo")
mean(SPX_SVICPV_zoo) # Note that that's the same as 'mean(SPX_dSVI_zoo)'; after much study on the matter, it seems as though this mean and other statistics are correct, which only highlights how different seperate SVI draws can be.
print("Standard Deviation of SPX_SVICPV_zoo")
sd(SPX_SVICPV_zoo)
print("Median of SPX_SVICPV_zoo")
median(SPX_SVICPV_zoo)
print("Skewness of SPX_SVICPV_zoo")
skewness(as.numeric(SPX_SVICPV_zoo))
print("Kurtosis of SPX_SVICPV_zoo")
kurtosis(as.numeric(SPX_SVICPV_zoo))
print("AutoCorrelation Function of SPX_SVICPV_zoo")
acf(as.numeric(SPX_SVICPV_zoo), lag.max = 1, plot = FALSE)
print("Augmented Dickey-Fuller Test of SPX_SVICPV_zoo")
adf.test(matrix(SPX_SVICPV_zoo), nlag = NULL, output = TRUE)
# Our ADF test here should show that we reject the Null Hypothesis (low p-value)
# in preference for the alternative of stationarity. One can see that dSVIFTSE_zoo
# is stationary by ploting it with:
plot(SPX_SVICPV_zoo, xlab='Date', ylab='SVICPVSPX', col='black', type='l',
# main='dSVIFTSE'
)
print("PP Test of SPX_SVICPV_zoo")
PP.test(SPX_SVICPV_zoo)
print("KPSS Test of SPX_SVICPV_zoo")
kpss.test(matrix(SPX_SVICPV_zoo))
# SPX's SPX_dSVICPV
print("Mean of SPX_SVICPV")
mean(SPX_dSVICPV)
print("Standard Deviation of SPX_SVICPV")
sd(SPX_dSVICPV)
print("Median of SPX_SVICPV")
median(SPX_dSVICPV)
print("Skewness of SPX_SVICPV")
skewness(as.numeric(SPX_dSVICPV))
print("Kurtosis of SPX_SVICPV")
kurtosis(as.numeric(SPX_dSVICPV))
print("AutoCorrelation Function of SPX_SVICPV")
acf(as.numeric(SPX_dSVICPV), lag.max = 1, plot = FALSE)
print("Augmented Dickey-Fuller Test of SPX_dSVICPV")
adf.test(matrix(SPX_dSVICPV), nlag = NULL, output = TRUE)
plot(SPX_dSVICPV, xlab='Date', ylab='dSVICPVSPX', col='black', type='l',
# main='dSVIFTSE'
)
print("PP Test of SPX_dSVICPV")
PP.test(SPX_dSVICPV)
print("KPSS Test of SPX_dSVICPV")
kpss.test(matrix(SPX_dSVICPV))
# SPX's SVICPVUS
print("Mean of SPX_SVICPVUS_zoo")
mean(SPX_SVICPVUS_zoo) # Note that that's the same as 'mean(SPX_dSVI_zoo)'; after much study on the matter, it seems as though this mean and other statistics are correct, which only highlights how different seperate SVI draws can be.
print("Standard Deviation of SPX_SVICPVUS_zoo")
sd(SPX_SVICPVUS_zoo)
print("Median of SPX_SVICPVUS_zoo")
median(SPX_SVICPVUS_zoo)
print("Skewness of SPX_SVICPVUS_zoo")
skewness(as.numeric(SPX_SVICPVUS_zoo))
print("Kurtosis of SPX_SVICPVUS_zoo")
kurtosis(as.numeric(SPX_SVICPVUS_zoo))
print("AutoCorrelation Function of SPX_SVICPVUS_zoo")
acf(as.numeric(SPX_SVICPVUS_zoo), lag.max = 1, plot = FALSE)
print("Augmented Dickey-Fuller Test of SPX_SVICPVUS_zoo")
adf.test(matrix(SPX_SVICPVUS_zoo), nlag = NULL, output = TRUE)
# Our ADF test here should show that we reject the Null Hypothesis (low p-value)
# in preference for the alternative of stationarity. One can see that dSVIFTSE_zoo
# is stationary by ploting it with:
plot(SPX_SVICPVUS_zoo, xlab='Date', ylab='SVICPVUSSPX', col='black', type='l',
# main='dSVIFTSE'
)
print("PP Test of SPX_SVICPVUS_zoo")
PP.test(SPX_SVICPVUS_zoo)
print("KPSS Test of SPX_SVICPVUS_zoo")
kpss.test(matrix(SPX_SVICPVUS_zoo))
# SPX's SPX_dSVICPVUS_all_days_zoo
print("Mean of SPX_dSVICPVUS_R")
mean(SPX_dSVICPVUS_R)
print("Standard Deviation of SPX_dSVICPVUS_R")
sd(SPX_dSVICPVUS_R)
print("Median of SPX_dSVICPVUS_R")
median(SPX_dSVICPVUS_R)
print("Skewness of SPX_dSVICPVUS_R")
skewness(as.numeric(SPX_dSVICPVUS_R))
print("Kurtosis of SPX_dSVICPVUS_R")
kurtosis(as.numeric(SPX_dSVICPVUS_R))
print("AutoCorrelation Function of SPX_dSVICPVUS_R")
acf(as.numeric(SPX_dSVICPVUS_R), lag.max = 1, plot = FALSE)
print("Augmented Dickey-Fuller Test of SPX_dSVICPVUS_R")
adf.test(matrix(SPX_dSVICPVUS_R), nlag = NULL, output = TRUE)
plot(SPX_dSVICPVUS_R, xlab='Date', ylab='dSVICPVUSSPX', col='black', type='l',
# main='dSVIFTSE'
)
print("PP Test of SPX_dSVICPVUS_R")
PP.test(SPX_dSVICPVUS_R)
print("KPSS Test of SPX_dSVICPVUS_R")
kpss.test(matrix(SPX_dSVICPVUS_R))
# Descriptive Plot
plot(SPX_SVI_zoo, xlab='Date', ylab='SVISPX1', col='blue', type='l',
# main='Google Search Vector Index 1 (SVI 1) for the SPX index over time'
)
plot(SPX_SVI2_zoo, xlab='Date', ylab='SVISPX2', col='blue',
# main='Google Search Vector Index 2 (SVI 2) for the SPX index over time', type='l'
)
plot(SPX_SVICPV_zoo, xlab='Date', ylab='SVISPXCPV', col='blue',
# main='Google Search Vector Index CPV (SVI-CPV) for the SPX index over time', type='l'
)
plot(SPX_SVICPVUS_zoo, xlab='Date', ylab='SVICPVUS', col='blue',
# main='Google Search Vector Index CPV-US (SVI-CPV-US) for the SPX index over time', type='l'
)
####---------------------------------------------------------------------------
#### FTSE GARCH models: create, train/fit them and create forecasts
####---------------------------------------------------------------------------
# Set parameters
roll = length(dSVIFTSE_zoo)-252
# This will also be the number of out-of-sample predictions. N.B.: the 253rd
# value of FTSE_R_zoo corresponds to 2005-01-04, the 1st trading day
# out-of-sample, the first value after 252.
# Define functions to make our code easier to read
GARCH_model_spec = function(mod, exreg)
{ugarchspec(variance.model = list(model=mod, garchOrder = c(1, 1),
external.regressors = exreg),
mean.model=list(armaOrder = c(1, 0), include.mean = TRUE,
external.regressors = exreg),
distribution.model = "norm")}
GARCH_model_forecast = function(mod)
{ugarchforecast(mod, n.ahead = 1, n.roll = 0, data = NULL, out.sample = 0)}
###----------------------------------------------------------------------------
### In-sample AR(1)-GARCH(1,1) Model
###---------------------------------------------------------------------------
in_sample_GARCH11 = GARCH_model_spec(mod="sGARCH", exreg= NULL)
in_sample_GARCH11fit = ugarchfit(data = FTSE_R_matrix[1:(251+1)],
spec=in_sample_GARCH11)
###----------------------------------------------------------------------------
### Create the AR(1)-GARCH(1,1) Model and forecasts
###----------------------------------------------------------------------------
# Non-linear Variance Models are created using the 'rugarch' package for
# adequacy; then create one-step-ahead forecasts, taking into account new
# data every step and re-estimating GARCH variables/cofactors.
GARCH11_mu_hat = zoo(matrix(, nrow=roll, ncol=1),
as.Date(Dates[(252+1):(252+roll)]))
colnames(GARCH11_mu_hat) = c('GARCH11_mu_hat')
GARCH11_sigma_hat = zoo(matrix(, nrow=roll, ncol=1),
as.Date(Dates[(252+1):(252+roll)]))
colnames(GARCH11_sigma_hat) = c('GARCH11_sigma_hat')
for (i in c(1:roll))
{GARCH11 = GARCH_model_spec(mod="sGARCH", exreg= NULL)
try(withTimeout({(GARCH11fit = ugarchfit(data = FTSE_R_matrix[1:(251+i)],
spec=GARCH11))}
, timeout = 3),silent = TRUE)
try((GARCH11fitforecast = GARCH_model_forecast(mod = GARCH11fit))
,silent = TRUE
# suppressWarnings(expr)
)
try((GARCH11_mu_hat[i]=GARCH11fitforecast@forecast[["seriesFor"]])
,silent = TRUE)
try((GARCH11_sigma_hat[i]=GARCH11fitforecast@forecast[["sigmaFor"]])
,silent = TRUE)
rm(GARCH11)
rm(GARCH11fit)
rm(GARCH11fitforecast)
}
# fitted(GARCH11fitforecast)
RV_no_GARCH11_sigma_hat_na_dated = as.matrix(RVFTSE_zoo
-(na.omit(GARCH11_sigma_hat*0)))
# This above is just a neat trick that returns strictly the RV (Realised
# Volatility) values corresponding to the dates for which GARCH11_sigma_hat
# values are not NA without any new functions or packages.
# Forecast Error, Mean and S.D.:
GARCH11forecasterror = (na.omit(GARCH11_sigma_hat)
-RV_no_GARCH11_sigma_hat_na_dated)
colnames(GARCH11forecasterror) = c("GARCH11forecasterror")
# plot(zoo(GARCH11forecasterror, as.Date(row.names(GARCH11forecasterror)))
# , type='l', ylab='GARCH11forecasterror', xlab='Date')
print("Mean of GARCH11forecasterror")
mean(GARCH11forecasterror)
print("Standard Deviation GARCH11forecasterror")
sd(GARCH11forecasterror)
# RMSE of the sigma (standard deviations) of the forecast:
print("RMSE of the sigma (standard deviations) of the forecast from the GARCH11 model")
rmse(RV_no_GARCH11_sigma_hat_na_dated, (na.omit(GARCH11_sigma_hat)))
# Note that this is the same as the bellow:
# sqrt(mean((RV[(252+1):(length(FTSE_R_matrix)+1)] -
# as.matrix((GARCH11fitforecast@forecast[["sigmaFor"]]))[1:roll])
# ^2))
###----------------------------------------------------------------------------
### In-sample AR(1)-GARCH(1,1)-SVI Model
###----------------------------------------------------------------------------
in_sample_GARCH11SVI = GARCH_model_spec(mod = "sGARCH", exreg = as.matrix(dSVIFTSE[1:(251+1)]))
in_sample_GARCH11SVIfit = ugarchfit(data = FTSE_R_matrix[1:(251+1)],
spec=in_sample_GARCH11SVI)
###----------------------------------------------------------------------------
### Create the AR(1)-GARCH(1,1)-SVI Model and forecasts
###----------------------------------------------------------------------------
GARCH11SVI_mu_hat = zoo(matrix(, nrow=roll, ncol=1), as.Date(Dates[(252+1):(252+roll)]))
colnames(GARCH11SVI_mu_hat) = c('GARCH11SVI_mu_hat')
GARCH11SVI_sigma_hat = zoo(matrix(, nrow=roll, ncol=1), as.Date(Dates[(252+1):(252+roll)]))
colnames(GARCH11SVI_sigma_hat) = c('GARCH11SVI_sigma_hat')
for (i in c(1:roll)){
GARCH11SVI = GARCH_model_spec(mod= "sGARCH", exreg = as.matrix(dSVIFTSE[1:(251+i)]))
try(withTimeout({(GARCH11SVIfit = ugarchfit(data = FTSE_R_matrix[1:(251+i)], spec=GARCH11SVI))}, timeout = 3),silent = TRUE)
try((GARCH11SVIfitforecast = GARCH_model_forecast(mod = GARCH11SVIfit)),silent = TRUE)
try((GARCH11SVI_mu_hat[i]=GARCH11SVIfitforecast@forecast[["seriesFor"]]), silent = TRUE)
try((GARCH11SVI_sigma_hat[i]=GARCH11SVIfitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(GARCH11SVI)
rm(GARCH11SVIfit)
rm(GARCH11SVIfitforecast)
}
# fitted(GARCH11SVIfitforecast)
RV_no_GARCH11SVI_sigma_hat_na_dated = as.matrix(RVFTSE_zoo
-(na.omit(GARCH11SVI_sigma_hat*0)))
# Forecast Error, Mean and S.D.:
GARCH11SVIforecasterror = (na.omit(GARCH11SVI_sigma_hat)
-RV_no_GARCH11SVI_sigma_hat_na_dated)
colnames(GARCH11SVIforecasterror) = c("GARCH11SVIforecasterror")
print("Mean of GARCH11SVIforecasterror")
mean(GARCH11SVIforecasterror)
print("Standard Deviation GARCH11SVIforecasterror")
sd(GARCH11SVIforecasterror)
# RMSE of the sigma (standard deviations) of the forecast:
print("RMSE of the sigma (standard deviations) of the forecast from the GARCH11-SVI model")
rmse(RV_no_GARCH11SVI_sigma_hat_na_dated, (na.omit(GARCH11SVI_sigma_hat)))
###----------------------------------------------------------------------------
### In-sample AR(1)-GJRGARCH(1,1) Model
###----------------------------------------------------------------------------
in_sample_GJRGARCH11 = GARCH_model_spec("gjrGARCH", exreg= NULL)
in_sample_GJRGARCH11fit = ugarchfit(data = FTSE_R_matrix[1:(251+1)], spec=in_sample_GJRGARCH11)
###----------------------------------------------------------------------------
### Create the AR(1)-GJRGARCH(1,1) Model and forecasts
###----------------------------------------------------------------------------
GJRGARCH11_mu_hat = zoo(matrix(, nrow=roll, ncol=1), as.Date(Dates[(252+1):(252+roll)]))
colnames(GJRGARCH11_mu_hat) = c('GJRGARCH11_mu_hat')
GJRGARCH11_sigma_hat = zoo(matrix(, nrow=roll, ncol=1), as.Date(Dates[(252+1):(252+roll)]))
colnames(GJRGARCH11_sigma_hat) = c('GJRGARCH11_sigma_hat')
for (i in c(1:roll))
{GJRGARCH11 = GARCH_model_spec(mod = "gjrGARCH", exreg = NULL)
try(withTimeout({(GJRGARCH11fit = ugarchfit(data = FTSE_R_matrix[1:(251+i)], spec=GJRGARCH11))}
, timeout = 3),silent = TRUE)
try((GJRGARCH11fitforecast = GARCH_model_forecast(mod = GJRGARCH11fit)),silent = TRUE)
try((GJRGARCH11_mu_hat[i]=GJRGARCH11fitforecast@forecast[["seriesFor"]]), silent = TRUE)
try((GJRGARCH11_sigma_hat[i]=GJRGARCH11fitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(GJRGARCH11)
rm(GJRGARCH11fit)
rm(GJRGARCH11fitforecast)
}
RV_no_GJRGARCH11_sigma_hat_na_dated = as.matrix(RVFTSE_zoo
-(na.omit(GJRGARCH11_sigma_hat*0)))
# Forecast Error, Mean and S.D.:
GJRGARCH11forecasterror = (na.omit(GJRGARCH11_sigma_hat)
-RV_no_GJRGARCH11_sigma_hat_na_dated)
colnames(GJRGARCH11forecasterror) = c("GJRGARCH11forecasterror")
print("Mean of GJRGARCH11forecasterror")
mean(GJRGARCH11forecasterror)
print("Standard Deviation GJRGARCH11forecasterror")
sd(GJRGARCH11forecasterror)
# RMSE of the sigma (standard deviations) of the forecast:
print("RMSE of the sigma (standard deviations) of the forecast from the GJRGARCH11 model")
rmse(RV_no_GJRGARCH11_sigma_hat_na_dated, (na.omit(GJRGARCH11_sigma_hat)))
###----------------------------------------------------------------------------
### In-sample AR(1)-GJRGARCH(1,1)-SVI Model
###----------------------------------------------------------------------------
in_sample_GJRGARCH11SVI = GARCH_model_spec("gjrGARCH", as.matrix(dSVIFTSE[1:(251+1)]))
in_sample_GJRGARCH11SVIfit = ugarchfit(data = FTSE_R_matrix[1:(251+1)],
spec=in_sample_GJRGARCH11SVI)
###----------------------------------------------------------------------------
### Create the AR(1)-GJRGARCH(1,1)-SVI Model and forecasts
###----------------------------------------------------------------------------
GJRGARCH11SVI_mu_hat = zoo(matrix(, nrow=roll, ncol=1),
as.Date(Dates[(252+1):(252+roll)]))
colnames(GJRGARCH11SVI_mu_hat) = c('GJRGARCH11SVI_mu_hat')
GJRGARCH11SVI_sigma_hat = zoo(matrix(, nrow=roll, ncol=1),
as.Date(Dates[(252+1):(252+roll)]))
colnames(GJRGARCH11SVI_sigma_hat) = c('GJRGARCH11SVI_sigma_hat')
for (i in c(1:roll))
{GJRGARCH11SVI = GARCH_model_spec(mod = "gjrGARCH", exreg=as.matrix(dSVIFTSE[1:(251+i)]))
try(withTimeout({(GJRGARCH11SVIfit = ugarchfit(data = FTSE_R_matrix[1:(251+i)],
spec=GJRGARCH11SVI))}
, timeout = 3),silent = TRUE)
try((GJRGARCH11SVIfitforecast = GARCH_model_forecast(mod = GJRGARCH11SVIfit)) ,silent = TRUE)
try((GJRGARCH11SVI_mu_hat[i]=GJRGARCH11SVIfitforecast@forecast[["seriesFor"]]), silent = TRUE)
try((GJRGARCH11SVI_sigma_hat[i]=GJRGARCH11SVIfitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(GJRGARCH11SVI)
rm(GJRGARCH11SVIfit)
rm(GJRGARCH11SVIfitforecast)
}
RV_no_GJRGARCH11SVI_sigma_hat_na_dated = as.matrix(RVFTSE_zoo
-(na.omit(GJRGARCH11SVI_sigma_hat*0)))
# Forecast Error, Mean and S.D.:
GJRGARCH11SVIforecasterror = (na.omit(GJRGARCH11SVI_sigma_hat)
-RV_no_GJRGARCH11SVI_sigma_hat_na_dated)
colnames(GJRGARCH11SVIforecasterror) = c("GJRGARCH11SVIforecasterror")
print("Mean of GJRGARCH11SVIforecasterror")
mean(GJRGARCH11SVIforecasterror)
print("Standard Deviation GJRGARCH11SVIforecasterror")
sd(GJRGARCH11SVIforecasterror)
# RMSE of the sigma (standard deviations) of the forecast:
print("RMSE of the sigma (standard deviations) of the forecast from the GJRGARCH11SVI model")
rmse(RV_no_GJRGARCH11SVI_sigma_hat_na_dated, (na.omit(GJRGARCH11SVI_sigma_hat)))
###----------------------------------------------------------------------------
### In-sample AR(1)-EGARCH(1,1) Model
###----------------------------------------------------------------------------
in_sample_EGARCH11 = GARCH_model_spec(mod="eGARCH", exreg= NULL)
in_sample_EGARCH11fit = ugarchfit(data = FTSE_R_matrix[1:(251+1)],
spec=in_sample_EGARCH11)
###----------------------------------------------------------------------------
### Create the AR(1)-EGARCH(1,1) Model and forecasts
###----------------------------------------------------------------------------
EGARCH11_mu_hat = zoo(matrix(, nrow=roll, ncol=1), as.Date(Dates[(252+1):(252+roll)]))
colnames(EGARCH11_mu_hat) = c('EGARCH11_mu_hat')
EGARCH11_sigma_hat = zoo(matrix(, nrow=roll, ncol=1), as.Date(Dates[(252+1):(252+roll)]))
colnames(EGARCH11_sigma_hat) = c('EGARCH11_sigma_hat')
for (i in c(1:roll))
{EGARCH11 = GARCH_model_spec(mod = "eGARCH", exreg = NULL)
try(withTimeout({(EGARCH11fit = ugarchfit(data = FTSE_R_matrix[1:(251+i)], spec=EGARCH11))}
, timeout = 3),silent = TRUE)
try((EGARCH11fitforecast = GARCH_model_forecast(mod = EGARCH11fit)),silent = TRUE)
try((EGARCH11_mu_hat[i]=EGARCH11fitforecast@forecast[["seriesFor"]]), silent = TRUE)
try((EGARCH11_sigma_hat[i]=EGARCH11fitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(EGARCH11)
rm(EGARCH11fit)
rm(EGARCH11fitforecast)
}
RV_no_EGARCH11_sigma_hat_na_dated = as.matrix(RVFTSE_zoo
-(na.omit(EGARCH11_sigma_hat*0)))
# Forecast Error, Mean and S.D.:
EGARCH11forecasterror = (na.omit(EGARCH11_sigma_hat)-RV_no_EGARCH11_sigma_hat_na_dated)
colnames(EGARCH11forecasterror) = c("EGARCH11forecasterror")
print("Mean of EGARCH11forecasterror")
mean(EGARCH11forecasterror)
print("Standard Deviation EGARCH11forecasterror")
sd(EGARCH11forecasterror)
# RMSE of the sigma (standard deviations) of the forecast:
print("RMSE of the sigma (standard deviations) of the forecast from the EGARCH11 model")
rmse(RV_no_EGARCH11_sigma_hat_na_dated, (na.omit(EGARCH11_sigma_hat)))
###----------------------------------------------------------------------------
### In-sample AR(1)-EGARCH(1,1)-SVI Model
###----------------------------------------------------------------------------
in_sample_EGARCH11SVI = GARCH_model_spec(mod = "eGARCH", exreg = as.matrix(dSVIFTSE[1:(251+1)]))
in_sample_EGARCH11SVIfit = ugarchfit(data = FTSE_R_matrix[1:(251+1)],
spec=in_sample_EGARCH11SVI)
###----------------------------------------------------------------------------
### Create the AR(1)-EGARCH(1,1)-SVI Model and forecasts
###----------------------------------------------------------------------------
EGARCH11SVI_mu_hat = zoo(matrix(, nrow=roll, ncol=1), as.Date(Dates[(252+1):(252+roll)]))
colnames(EGARCH11SVI_mu_hat) = c('EGARCH11SVI_mu_hat')
EGARCH11SVI_sigma_hat = zoo(matrix(, nrow=roll, ncol=1), as.Date(Dates[(252+1):(252+roll)]))
colnames(EGARCH11SVI_sigma_hat) = c('EGARCH11SVI_sigma_hat')
for (i in c(1:roll))
{EGARCH11SVI = GARCH_model_spec(mod = "eGARCH", exreg = as.matrix(dSVIFTSE[1:(251+i)]))
try(withTimeout({(EGARCH11SVIfit = ugarchfit(data = FTSE_R_matrix[1:(251+i)], spec=EGARCH11SVI))}
, timeout = 3),silent = TRUE)
try((EGARCH11SVIfitforecast = GARCH_model_forecast(mod = EGARCH11SVIfit)), silent = TRUE)
try((EGARCH11SVI_mu_hat[i]=EGARCH11SVIfitforecast@forecast[["seriesFor"]]), silent = TRUE)
try((EGARCH11SVI_sigma_hat[i]=EGARCH11SVIfitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(EGARCH11SVI)
rm(EGARCH11SVIfit)
rm(EGARCH11SVIfitforecast)
}
# fitted(EGARCH11SVIfitforecast)
RV_no_EGARCH11SVI_sigma_hat_na_dated = as.matrix(RVFTSE_zoo
-(na.omit(EGARCH11SVI_sigma_hat*0)))
# Forecast Error, Mean and S.D.:
EGARCH11SVIforecasterror = (na.omit(EGARCH11SVI_sigma_hat)-RV_no_EGARCH11SVI_sigma_hat_na_dated)
colnames(EGARCH11SVIforecasterror) = c("EGARCH11SVIforecasterror")
print("Mean of EGARCH11SVIforecasterror")
mean(EGARCH11SVIforecasterror)
print("Standard Deviation EGARCH11SVIforecasterror")
sd(EGARCH11SVIforecasterror)
# RMSE of the sigma (standard deviations) of the forecast:
print("RMSE of the sigma (standard deviations) of the forecast from the EGARCH11-SVI model")
rmse(RV_no_EGARCH11SVI_sigma_hat_na_dated, (na.omit(EGARCH11SVI_sigma_hat)))
###---------------------------------------------------------------------------
### s.d. forecase D-M tests
###---------------------------------------------------------------------------
print("This function implements the modified test proposed by Harvey, Leybourne and Newbold (1997). The null hypothesis is that the two methods have the same forecast accuracy. For alternative=less, the alternative hypothesis is that method 2 is less accurate than method 1. For alternative=greater, the alternative hypothesis is that method 2 is more accurate than method 1. For alternative=two.sided, the alternative hypothesis is that method 1 and method 2 have different levels of accuracy.")
print("Diebold and Mariano test GARCH11 with and without SVI")
GARCH11forecasterror_dm = GARCH11forecasterror - (GARCH11SVIforecasterror*0)
GARCH11SVIforecasterror_dm = GARCH11SVIforecasterror - (GARCH11forecasterror*0)
dm.test(matrix(GARCH11forecasterror_dm), matrix(GARCH11SVIforecasterror_dm), alternative = "less")
print("Diebold and Mariano test GJRGARCH11 with and without SVI")
GJRGARCH11forecasterror_dm = GJRGARCH11forecasterror - (GJRGARCH11SVIforecasterror*0)
GJRGARCH11SVIforecasterror_dm = GJRGARCH11SVIforecasterror - (GJRGARCH11forecasterror*0)
dm.test(matrix(GJRGARCH11forecasterror_dm), matrix(GJRGARCH11SVIforecasterror_dm), alternative = "less")
print("Diebold and Mariano test EGARCH11 with and without SVI")
EGARCH11forecasterror_dm = EGARCH11forecasterror - (EGARCH11SVIforecasterror*0)
EGARCH11SVIforecasterror_dm = EGARCH11SVIforecasterror - (EGARCH11forecasterror*0)
dm.test(matrix(EGARCH11forecasterror_dm), matrix(EGARCH11SVIforecasterror_dm), alternative = "less")
####---------------------------------------------------------------------------
#### FTSE estimates of the probability of a positive return
####---------------------------------------------------------------------------
###----------------------------------------------------------------------------
### C&D, Naive
###----------------------------------------------------------------------------
##-----------------------------------------------------------------------------
## Naive-Model and its forecast error statistics
##-----------------------------------------------------------------------------
# Naive-Model's Indicator function according to the GARCH11 Model:
Naive_I = ifelse(FTSE_R_matrix>0 , 1 , 0)
# Naive Model:
Naive=(1:(length(FTSE_R_matrix)))
for(t in c(1:(length(FTSE_R_matrix))))
{Naive[t] = (1/t) * sum(FTSE_R_matrix[1:t])}
# Note that the naive model provides the probability of aa positive return in
# the next time period and therefore spams from 2004-01-06 to 2019-03-13.
Naive_zoo = zoo(Naive, as.Date(rownames(as.matrix(FTSE_R_zoo))))
##-----------------------------------------------------------------------------
## C&D's GARCH models with and without SVI
## (Model independent variables' construction)
##-----------------------------------------------------------------------------
# mean of return up to time 'k'
R_mu=(1:length(FTSE_R_matrix))
for (i in c(1:length(FTSE_R_matrix))){R_mu[i]=mean(FTSE_R_matrix[1:i])}
# standard deviation of return up to time 'k'.
# Note that its 1st value is NA since there is no standard deviation for a
# constant (according to R).
R_sigma=(1:length(FTSE_R_matrix))
for (i in c(1:length(FTSE_R_matrix))){R_sigma[i]=sd(FTSE_R_matrix[1:i])}
R_sigma[1]=0 # Since the variance of a constant is 0.
#------------------------------------------------------------------------------
# GARCH11's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GARCH11 Model:
GARCH11_CandD_I = matrix(, nrow=roll, ncol=roll)
GARCH11_CandD_I_k= matrix(, nrow=roll, ncol=roll)
GARCH11_CandD_I_t = matrix( - (GARCH11_mu_hat/GARCH11_sigma_hat))
for(t in c(1:roll))
{for (k in c(1:t))
{GARCH11_CandD_I_k[k,t] =
ifelse(((FTSE_R_matrix[252+k]-GARCH11_mu_hat[k])/GARCH11_sigma_hat[k])
<=GARCH11_CandD_I_t[t+1],
1,0)
# Note that we have some missing values due to the model not always managing
# to converge to estimates of sigma and mu; since we need the t+1 model
# estimates to compute its CandD_I_k, we also loose its t'th value for
# each model's missing value.
# We also los the last value (the roll'th value) for the same reason.
GARCH11_CandD_I[k,t] =
ifelse((is.na(GARCH11_CandD_I_k[k,t])), NA,
(sum(GARCH11_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GARCH11 Model:
GARCH11_CandD=(1:(roll-1))
for(i in c(1:(roll-1)))
{GARCH11_CandD[i] = 1 - ((GARCH11_CandD_I[i,i])/
length(na.omit(GARCH11_CandD_I[1:i,i])))}
GARCH11_CandD_zoo = zoo(GARCH11_CandD,
as.Date(Dates[(253+1):(253+(roll-1))]))
# This zoo object has missing values.
GARCH11_CandD_zoo_no_nas=zoo(na.omit(GARCH11_CandD_zoo))
# rm(GARCH11_CandD_I_t) # Cleaning datasets
# rm(GARCH11_CandD_I_k)
# rm(GARCH11_CandD_I)
#------------------------------------------------------------------------------
# GARCH11-SVI's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GARCH11SVI Model:
GARCH11SVI_CandD_I = matrix(, nrow=roll, ncol=roll)
GARCH11SVI_CandD_I_k= matrix(, nrow=roll, ncol=roll)
GARCH11SVI_CandD_I_t = matrix( - (GARCH11SVI_mu_hat/GARCH11SVI_sigma_hat))
for(t in c(1:roll))
{for (k in c(1:t))
{GARCH11SVI_CandD_I_k[k,t] =
ifelse(((FTSE_R_matrix[252+k]-GARCH11SVI_mu_hat[k])/GARCH11SVI_sigma_hat[k])
<=GARCH11SVI_CandD_I_t[t+1],
1,0)
GARCH11SVI_CandD_I[k,t] =
ifelse(
(is.na(GARCH11SVI_CandD_I_k[k,t])), NA,
(sum(GARCH11SVI_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GARCH11SVI Model:
GARCH11SVI_CandD=(1:(roll-1))
for(i in c(1:(roll-1)))
{GARCH11SVI_CandD[i] = 1 - ((GARCH11SVI_CandD_I[i,i])/
length(na.omit(GARCH11SVI_CandD_I[1:i,i])))}
GARCH11SVI_CandD_zoo = zoo(GARCH11SVI_CandD,
as.Date(Dates[(253+1):(253+(roll-1))]))
# This zoo object has missing values.
GARCH11SVI_CandD_zoo_no_nas=zoo(na.omit(GARCH11SVI_CandD_zoo))
# rm(GARCH11SVI_CandD_I_t) # Cleaning datasets
# rm(GARCH11SVI_CandD_I_k)
# rm(GARCH11SVI_CandD_I)
#------------------------------------------------------------------------------
# GJRGARCH11's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GJRGARCH11 Model:
GJRGARCH11_CandD_I = matrix(, nrow=roll, ncol=roll)
GJRGARCH11_CandD_I_k= matrix(, nrow=roll, ncol=roll)
GJRGARCH11_CandD_I_t = matrix( - (GJRGARCH11_mu_hat/GJRGARCH11_sigma_hat))
for(t in c(1:roll))
{for (k in c(1:t))
{GJRGARCH11_CandD_I_k[k,t] =
ifelse(((FTSE_R_matrix[252+k]-GJRGARCH11_mu_hat[k])/GJRGARCH11_sigma_hat[k])
<=GJRGARCH11_CandD_I_t[t+1],
1,0)
# Note that we have some missing values due to the model not always managing
# to converge to estimates of sigma and mu; since we need the t+1 model
# estimates to compute its CandD_I_k, we also loose its t'th value for
# each model's missing value.
# We also los the last value (the roll'th value) for the same reason.
GJRGARCH11_CandD_I[k,t] =
ifelse(
(is.na(GJRGARCH11_CandD_I_k[k,t])), NA,
(sum(GJRGARCH11_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GJRGARCH11 Model:
GJRGARCH11_CandD=(1:(roll-1))
for(i in c(1:(roll-1)))
{GJRGARCH11_CandD[i] = 1 - ((GJRGARCH11_CandD_I[i,i])/
length(na.omit(GJRGARCH11_CandD_I[1:i,i])))}
GJRGARCH11_CandD_zoo = zoo(GJRGARCH11_CandD,
as.Date(Dates[(253+1):(253+(roll-1))]))
# This zoo object has missing values.
GJRGARCH11_CandD_zoo_no_nas=zoo(na.omit(GJRGARCH11_CandD_zoo))
# rm(GJRGARCH11_CandD_I_t) # Cleaning datasets
# rm(GJRGARCH11_CandD_I_k)
# rm(GJRGARCH11_CandD_I)
#------------------------------------------------------------------------------
# GJRGARCH11-SVI's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GJRGARCH11-SVI Model:
GJRGARCH11SVI_CandD_I = matrix(, nrow=roll, ncol=roll)
GJRGARCH11SVI_CandD_I_k = matrix(, nrow=roll, ncol=roll)
GJRGARCH11SVI_CandD_I_t = matrix( - (GJRGARCH11SVI_mu_hat/GJRGARCH11SVI_sigma_hat))
for(t in c(1:roll))
{for (k in c(1:t))
{GJRGARCH11SVI_CandD_I_k[k,t] =
ifelse(((FTSE_R_matrix[252+k]-GJRGARCH11SVI_mu_hat[k])/GJRGARCH11SVI_sigma_hat[k])
<=GJRGARCH11SVI_CandD_I_t[t+1],
1,0)
GJRGARCH11SVI_CandD_I[k,t] =
ifelse(
(is.na(GJRGARCH11SVI_CandD_I_k[k,t])), NA,
(sum(GJRGARCH11SVI_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GJRGARCH11-SVI Model:
GJRGARCH11SVI_CandD=(1:(roll-1))
for(i in c(1:(roll-1)))
{GJRGARCH11SVI_CandD[i] = 1 - ((GJRGARCH11SVI_CandD_I[i,i])/
length(na.omit(GJRGARCH11SVI_CandD_I[1:i,i])))}
GJRGARCH11SVI_CandD_zoo = zoo(GJRGARCH11SVI_CandD,
as.Date(Dates[(253+1):(253+(roll-1))]))
# This zoo object has missing values.
GJRGARCH11SVI_CandD_zoo_no_nas=zoo(na.omit(GJRGARCH11SVI_CandD_zoo))
# rm(GJRGARCH11SVI_CandD_I_t) # Cleaning datasets
# rm(GJRGARCH11SVI_CandD_I_k)
# rm(GJRGARCH11SVI_CandD_I)
#------------------------------------------------------------------------------
# EGARCH11's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the EGARCH11 Model:
EGARCH11_CandD_I = matrix(, nrow=roll, ncol=roll)
EGARCH11_CandD_I_k= matrix(, nrow=roll, ncol=roll)
EGARCH11_CandD_I_t = matrix( - (EGARCH11_mu_hat/EGARCH11_sigma_hat))
for(t in c(1:roll))
{for (k in c(1:t))
{EGARCH11_CandD_I_k[k,t] =
ifelse(((FTSE_R_matrix[252+k]-EGARCH11_mu_hat[k])/EGARCH11_sigma_hat[k])
<=EGARCH11_CandD_I_t[t+1],
1,0)
EGARCH11_CandD_I[k,t] =
ifelse(
(is.na(EGARCH11_CandD_I_k[k,t])), NA,
(sum(EGARCH11_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the EGARCH11 Model:
EGARCH11_CandD=(1:(roll-1))
for(i in c(1:(roll-1)))
{EGARCH11_CandD[i] = 1 - ((EGARCH11_CandD_I[i,i])/
length(na.omit(EGARCH11_CandD_I[1:i,i])))}
EGARCH11_CandD_zoo = zoo(EGARCH11_CandD,
as.Date(Dates[(253+1):(253+(roll-1))]))
# This zoo object has missing values.
EGARCH11_CandD_zoo_no_nas=zoo(na.omit(EGARCH11_CandD_zoo))
# rm(EGARCH11_CandD_I_t) # Cleaning datasets
# rm(EGARCH11_CandD_I_k)
# rm(EGARCH11_CandD_I)
#------------------------------------------------------------------------------
# EGARCH11-SVI's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the EGARCH11-SVI Model:
EGARCH11SVI_CandD_I = matrix(, nrow=roll, ncol=roll)
EGARCH11SVI_CandD_I_k= matrix(, nrow=roll, ncol=roll)
EGARCH11SVI_CandD_I_t = matrix( - (EGARCH11SVI_mu_hat/EGARCH11SVI_sigma_hat))
for(t in c(1:roll))
{for (k in c(1:t))
{EGARCH11SVI_CandD_I_k[k,t] =
ifelse(((FTSE_R_matrix[252+k]-EGARCH11SVI_mu_hat[k])/EGARCH11SVI_sigma_hat[k])
<=EGARCH11SVI_CandD_I_t[t+1],
1,0)
EGARCH11SVI_CandD_I[k,t] =
ifelse(
(is.na(EGARCH11SVI_CandD_I_k[k,t])), NA,
(sum(EGARCH11SVI_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the EGARCH11-SVI Model:
EGARCH11SVI_CandD=(1:(roll-1))
for(i in c(1:(roll-1)))
{EGARCH11SVI_CandD[i] = 1 - ((EGARCH11SVI_CandD_I[i,i])/
length(na.omit(EGARCH11SVI_CandD_I[1:i,i])))}
EGARCH11SVI_CandD_zoo = zoo(EGARCH11SVI_CandD,
as.Date(Dates[(253+1):(253+(roll-1))]))
# This zoo object has missing values.
EGARCH11SVI_CandD_zoo_no_nas=zoo(na.omit(EGARCH11SVI_CandD_zoo))
# rm(EGARCH11SVI_CandD_I_t) # Cleaning datasets
# rm(EGARCH11SVI_CandD_I_k)
# rm(EGARCH11SVI_CandD_I)
##-----------------------------------------------------------------------------
## compare the predictive performance of each model
## (forecast error mean, sd, Brier Scores and Diebold&Mariano statistics)
##-----------------------------------------------------------------------------
# # mean of the probabilistic forecast errors derived from each model
Observed_pi= ifelse(FTSE_R_matrix>0,1,0)
Observed_pi_zoo = zoo(Observed_pi, as.Date(Dates))
Naive_pi_error = Naive_zoo - Observed_pi_zoo
GARCH11_pi_error = GARCH11_CandD_zoo_no_nas - Observed_pi_zoo
GARCH11SVI_pi_error = GARCH11SVI_CandD_zoo_no_nas - Observed_pi_zoo
GJRGARCH11_pi_error = GJRGARCH11_CandD_zoo_no_nas - Observed_pi_zoo
GJRGARCH11SVI_pi_error = GJRGARCH11SVI_CandD_zoo_no_nas - Observed_pi_zoo
EGARCH11_pi_error = EGARCH11_CandD_zoo_no_nas - Observed_pi_zoo
EGARCH11SVI_pi_error = EGARCH11SVI_CandD_zoo_no_nas - Observed_pi_zoo
mean(Naive_pi_error)
sd(Naive_pi_error)
mean(GARCH11_pi_error)
sd(GARCH11_pi_error)
mean(GARCH11SVI_pi_error)
sd(GARCH11SVI_pi_error)
mean(GJRGARCH11_pi_error)
sd(GJRGARCH11_pi_error)
mean(GJRGARCH11SVI_pi_error)
sd(GJRGARCH11SVI_pi_error)
mean(EGARCH11_pi_error)
sd(EGARCH11_pi_error)
mean(EGARCH11SVI_pi_error)
sd(EGARCH11SVI_pi_error)
# # Brier scores of the probabilistic forecast errors derived from each model
Naive_pi_error_Brier_score =
(1/length(Naive_pi_error))*sum(Naive_pi_error^2)
show(Naive_pi_error_Brier_score)
GARCH11_pi_error_Brier_score =
(1/length(GARCH11_pi_error))*sum(GARCH11_pi_error^2)
show(GARCH11_pi_error_Brier_score)
GARCH11SVI_pi_error_Brier_score =
(1/length(GARCH11SVI_pi_error))*sum(GARCH11SVI_pi_error^2)
show(GARCH11SVI_pi_error_Brier_score)
GJRGARCH11_pi_error_Brier_score =
(1/length(GJRGARCH11_pi_error))*sum(GJRGARCH11_pi_error^2)
show(GJRGARCH11_pi_error_Brier_score)
GJRGARCH11SVI_pi_error_Brier_score =
(1/length(GJRGARCH11SVI_pi_error))*sum(GJRGARCH11SVI_pi_error^2)
show(GJRGARCH11SVI_pi_error_Brier_score)
EGARCH11_pi_error_Brier_score =
(1/length(EGARCH11_pi_error))*sum(EGARCH11_pi_error^2)
show(EGARCH11_pi_error_Brier_score)
EGARCH11SVI_pi_error_Brier_score =
(1/length(EGARCH11SVI_pi_error))*sum(EGARCH11SVI_pi_error^2)
show(EGARCH11SVI_pi_error_Brier_score)
# # Diebold&Mariano statistics
# Here the alternative hypothesis is that method 2 is more accurate than
# method 1; remember that a small p-value indicates strong evidence against
# the Null Hypothesis.
GARCH11_pi_error_dm = GARCH11_pi_error - (GARCH11SVI_pi_error*0)
GARCH11SVI_pi_error_dm = GARCH11SVI_pi_error - (GARCH11_pi_error*0)
dm.test(matrix(GARCH11_pi_error_dm),
matrix(GARCH11SVI_pi_error_dm), alternative = "less")
GJRGARCH11_pi_error_dm = GJRGARCH11_pi_error - (GJRGARCH11SVI_pi_error*0)
GJRGARCH11SVI_pi_error_dm = GJRGARCH11SVI_pi_error - (GJRGARCH11_pi_error*0)
dm.test(matrix(GJRGARCH11_pi_error_dm),
matrix(GJRGARCH11SVI_pi_error_dm), alternative = c("greater"))
EGARCH11_pi_error_dm = EGARCH11_pi_error - (EGARCH11SVI_pi_error*0)
EGARCH11SVI_pi_error_dm = EGARCH11SVI_pi_error - (EGARCH11_pi_error*0)
dm.test(matrix(EGARCH11_pi_error_dm),
matrix(EGARCH11SVI_pi_error_dm), alternative = c("greater"))
####----------------------------------------------------------------------------
#### FTSE Financial significance
####----------------------------------------------------------------------------
# As per Chronopoulos et. al (2018): Investor with a utility function for wealth w is defined as U(w) where
U = function(w) {-exp((-3) * w)}
# where A is the investor's degree of risk aversion.
# in Goetzmann et al. (2007), Della Corte et al. (2010), and Andriosopoulos et al. (2014), we assume that the risk aversion coefficient to be 3.
# indicator of the realised direction of the return on the S&P 500 index
y_d = ifelse(FTSE_R_zoo>0,1,0)
# Set the probability threashold at which to invest in the index in our 2d graphs:
p = 0.5
###----------------------------------------------------------------------------
### FTSE Granger and Pesaran (2000)'s framework using the Naive model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_Naive = ifelse(Naive_zoo>p,1,0)
y_d_Naive = y_d - (Naive_zoo*0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_Naive),1))==1 , 1, 0)
R_Active_Naive_p = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_Naive*0))+(BoE_r_f_zoo-(y_hat_d_Naive*0))) +
(1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_Naive*0))))[2:length(y_hat_d_Naive)]
# note that the trick '(FTSE_R_zoo-(y_hat_d_Naive*0))' returns only the R values
# corresponding to dates present in y_hat_d_Naive, which are the ones
# in omega (omega couldn't be used because it's not a zoo object)
R_Active_Naive_p_cumulated = matrix(,nrow=length(R_Active_Naive_p))
for (i in c(1:length(R_Active_Naive_p)))
{R_Active_Naive_p_cumulated[i] = prod((1+R_Active_Naive_p)[1:i])}
# plot(R_Active_Naive_p_cumulated, type='l')
R_cumulated = matrix(,nrow=length(R_Active_Naive_p))
for (i in c(1:length(R_Active_Naive_p)))
{R_cumulated[i] = prod((1+(FTSE_R_zoo+(BoE_r_f_zoo - (R_Active_Naive_p*0))))[1:i])}
# plot(R_cumulated, type='l')
plot(zoo(R_cumulated,
as.Date(zoo::index(y_hat_d_Naive[2:length(y_hat_d_Naive)]))),
type="l",col="black", xlab='Date',
ylab='FTSE index and Naive index cumulated')
lines(zoo(R_Active_Naive_p_cumulated,
as.Date(zoo::index(y_hat_d_Naive[2:length(y_hat_d_Naive)]))),
col="green")
###----------------------------------------------------------------------------
### Granger and Pesaran (2000)'s framework using the GARCH11 model
###----------------------------------------------------------------------------
##----------------------------------------------------------------------------
## FTSE 3D graph Low Resolution
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
sequance = seq(from = 0, to = 1, by = 0.05)
R_Active_GARCH11_p = matrix(, nrow=3575, ncol=length(sequance))
R_Active_GARCH11_p_cumulated = matrix(, nrow=3575, ncol=length(sequance))
R_cumulated = matrix(, nrow=3575, ncol=length(sequance))
y_d_GARCH11 = y_d - (GARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GARCH11's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(length(sequance)-1)
# corresponding directional forecast and realised direction
y_hat_d_GARCH11 = ifelse(GARCH11_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_GARCH11),1))==1 , 1, 0)
R_Active_GARCH11_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_GARCH11*0))+(BoE_r_f_zoo-(y_hat_d_GARCH11*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_GARCH11*0))))[2:length(y_hat_d_GARCH11)] # note that the trick '(FTSE_R_zoo-(y_hat_d_GARCH11*0))' returns only the R values corresponding to dates present in y_hat_d_GARCH11, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_GARCH11_p[,P_number] = R_Active_GARCH11_p_col_zoo
for (i in c(1:length(R_Active_GARCH11_p[,P_number])))
{R_Active_GARCH11_p_cumulated[i,P_number] = prod((1+R_Active_GARCH11_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_GARCH11_p_col_zoo*0)))[1:i])}
}
GARCH11_CandD_3D_LowRes = (plot_ly(z=R_Active_GARCH11_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using GARCH11 models",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(GARCH11_CandD_3D, filename = "GARCH11_CandD_3D-public-graph")
##----------------------------------------------------------------------------
## 3D graph High Resolution
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
BY = 0.0125
FROM = 0.45
sequance = seq(from = FROM, to = 0.6, by = BY)
R_Active_GARCH11_p = matrix(, nrow=3575, ncol=length(sequance))
R_Active_GARCH11_p_cumulated = matrix(, nrow=3575, ncol=length(sequance))
R_cumulated = matrix(, nrow=3575, ncol=length(sequance))
y_d_GARCH11 = y_d - (GARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GARCH11's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(1/BY) - FROM/BY
# corresponding directional forecast and realised direction
y_hat_d_GARCH11 = ifelse(GARCH11_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_GARCH11),1))==1 , 1, 0)
R_Active_GARCH11_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_GARCH11*0))+ (BoE_r_f_zoo-(y_hat_d_GARCH11*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_GARCH11*0))))[2:length(y_hat_d_GARCH11)] # note that the trick '(FTSE_R_zoo-(y_hat_d_GARCH11*0))' returns only the R values corresponding to dates present in y_hat_d_GARCH11, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_GARCH11_p[,P_number] = R_Active_GARCH11_p_col_zoo
for (i in c(1:length(R_Active_GARCH11_p[,P_number])))
{R_Active_GARCH11_p_cumulated[i,P_number] = prod((1+R_Active_GARCH11_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_GARCH11_p_col_zoo*0)))[1:i])}
}
GARCH11_CandD_3D_HighRes = (plot_ly(z=R_Active_GARCH11_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using GARCH11 models where 0.45<=P<=0.6",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(GARCH11_CandD_3D, filename = "GARCH11_CandD_3D-public-graph")
##----------------------------------------------------------------------------
## 2D graph (p=0.49 seems best at increments of 0.0005 p increpemnts)
##----------------------------------------------------------------------------
p = 0.49
# corresponding directional forecast and realised direction
y_hat_d_GARCH11_2d = ifelse(GARCH11_CandD_zoo_no_nas>p,1,0)
y_d_GARCH11_2d = y_d - (GARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GARCH11-'s.
# let the portfolio weight attributed to the stock market index be
omega_2d = ifelse( (lead(as.matrix(y_hat_d_GARCH11_2d),1))==1 , 1, 0)
R_Active_GARCH11_p_2d = ((lag(omega_2d,1) * (FTSE_R_zoo-(y_hat_d_GARCH11_2d*0)) +
(1-lag(omega_2d,1)) * (BoE_r_f_zoo-(y_hat_d_GARCH11_2d*0))))[2:length(y_hat_d_GARCH11)]
# note that the trick '(FTSE_R_zoo-(y_hat_d_GARCH11_2d*0))' returns only the R values
# corresponding to dates present in y_hat_d_GARCH11_2d, which are the ones
# in omega (omega_2d couldn't be used because it's not a zoo object)
R_Active_GARCH11_p_cumulated_2d = matrix(,nrow=length(R_Active_GARCH11_p_2d))
for (i in c(1:length(R_Active_GARCH11_p_2d)))
{R_Active_GARCH11_p_cumulated_2d[i] = prod((1+R_Active_GARCH11_p_2d)[1:i])}
# plot(R_Active_GARCH11_p_cumulated_2d, type='l')
R_cumulated_2d = matrix(,nrow=length(R_Active_GARCH11_p_2d))
for (i in c(1:length(R_Active_GARCH11_p_2d)))
{R_cumulated_2d[i] = prod((1+(FTSE_R_zoo+(BoE_r_f_zoo - (R_Active_GARCH11_p_2d*0))))[1:i])}
# plot(R_cumulated_2d, type='l')
plot(zoo(R_cumulated_2d,
as.Date(zoo::index(y_hat_d_GARCH11_2d[2:length(y_hat_d_GARCH11_2d)]))),
type="l",col="black", xlab='Date', #ylim=c(0.5,1.2),
ylab=p)
lines(zoo(R_Active_GARCH11_p_cumulated_2d,
as.Date(zoo::index(y_hat_d_GARCH11_2d[2:length(y_hat_d_GARCH11_2d)]))),
col="red")
lines(zoo(R_Active_Naive_p_cumulated,
as.Date(zoo::index(y_hat_d_Naive[2:length(y_hat_d_Naive)]))),
col="green")
###----------------------------------------------------------------------------
### FTSE Granger and Pesaran (2000)'s framework using the GARCH11-SVI model (p=0.4935 seems best)
###----------------------------------------------------------------------------
p=0.4935
# corresponding directional forecast and realised direction
y_hat_d_GARCH11SVI = ifelse(GARCH11SVI_CandD_zoo_no_nas>p,1,0)
y_d_GARCH11SVI = y_d - (GARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has
# values on dates corresponding to GARCH11-SVI's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_GARCH11SVI),1))==1 , 1, 0)
R_Active_GARCH11SVI_p = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_GARCH11SVI*0))+(BoE_r_f_zoo-(y_hat_d_GARCH11SVI*0))) +
(1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_GARCH11SVI*0))))[2:length(y_hat_d_GARCH11SVI)]
# note that the trick '(FTSE_R_zoo-(y_hat_d_GARCH11SVI*0))' returns only the R values
# corresponding to dates present in y_hat_d_GARCH11SVI, which are the ones
# in omega (omega couldn't be used because it's not a zoo object)
R_Active_GARCH11SVI_p_cumulated = matrix(,nrow=length(R_Active_GARCH11SVI_p))
for (i in c(1:length(R_Active_GARCH11SVI_p)))
{R_Active_GARCH11SVI_p_cumulated[i] = prod((1+R_Active_GARCH11SVI_p)[1:i])}
# plot(R_Active_GARCH11SVI_p_cumulated, type='l')
R_cumulated = matrix(,nrow=length(R_Active_GARCH11SVI_p))
for (i in c(1:length(R_Active_GARCH11SVI_p)))
{R_cumulated[i] = prod((1+(FTSE_R_zoo+(BoE_r_f_zoo - (R_Active_GARCH11SVI_p*0))))[1:i])}
# plot(R_cumulated, type='l')
plot(zoo(R_cumulated,
as.Date(zoo::index(y_hat_d_GARCH11SVI[2:length(y_hat_d_GARCH11SVI)]))),
type="l",col="black", xlab='Date', ylim=c(0.67,2), main=p,
ylab='FTSE index and GARCH11SVI index cumulated')
lines(zoo(R_Active_GARCH11SVI_p_cumulated,
as.Date(zoo::index(y_hat_d_GARCH11SVI[2:length(y_hat_d_GARCH11SVI)]))),
col="green")
##----------------------------------------------------------------------------
## 3D graph Low Resolution
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
sequance = seq(from = 0, to = 1, by = 0.05)
R_Active_GARCH11SVI_p = matrix(, nrow=3567, ncol=length(sequance))
R_Active_GARCH11SVI_p_cumulated = matrix(, nrow=3567, ncol=length(sequance))
R_cumulated = matrix(, nrow=3567, ncol=length(sequance))
y_d_GARCH11SVI = y_d - (GARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GARCH11SVI's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(length(sequance)-1)
# corresponding directional forecast and realised direction
y_hat_d_GARCH11SVI = ifelse(GARCH11SVI_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_GARCH11SVI),1))==1 , 1, 0)
R_Active_GARCH11SVI_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_GARCH11SVI*0))+(BoE_r_f_zoo-(y_hat_d_GARCH11SVI*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_GARCH11SVI*0))))[2:length(y_hat_d_GARCH11SVI)] # note that the trick '(FTSE_R_zoo-(y_hat_d_GARCH11SVI*0))' returns only the R values corresponding to dates present in y_hat_d_GARCH11SVI, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_GARCH11SVI_p[,P_number] = R_Active_GARCH11SVI_p_col_zoo
for (i in c(1:length(R_Active_GARCH11SVI_p[,P_number])))
{R_Active_GARCH11SVI_p_cumulated[i,P_number] = prod((1+R_Active_GARCH11SVI_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_GARCH11SVI_p_col_zoo*0)))[1:i])}
}
GARCH11SVI_CandD_3D_LowRes = (plot_ly(z=R_Active_GARCH11SVI_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using GARCH11SVI models",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(GARCH11SVI_CandD_3D, filename = "GARCH11SVI_CandD_3D-public-graph")
##----------------------------------------------------------------------------
## 3D graph High Resolution (suggests P =0.45)
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
BY = 0.0125
FROM = 0.4
sequance = seq(from = FROM, to = 0.55, by = BY)
R_Active_GARCH11SVI_p = matrix(, nrow=3567, ncol=length(sequance))
R_Active_GARCH11SVI_p_cumulated = matrix(, nrow=3567, ncol=length(sequance))
R_cumulated = matrix(, nrow=3567, ncol=length(sequance))
y_d_GARCH11SVI = y_d - (GARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GARCH11SVI's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(1/BY) - FROM/BY
# corresponding directional forecast and realised direction
y_hat_d_GARCH11SVI = ifelse(GARCH11SVI_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_GARCH11SVI),1))==1 , 1, 0)
R_Active_GARCH11SVI_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_GARCH11SVI*0))+ (BoE_r_f_zoo-(y_hat_d_GARCH11SVI*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_GARCH11SVI*0))))[2:length(y_hat_d_GARCH11SVI)] # note that the trick '(FTSE_R_zoo-(y_hat_d_GARCH11SVI*0))' returns only the R values corresponding to dates present in y_hat_d_GARCH11SVI, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_GARCH11SVI_p[,P_number] = R_Active_GARCH11SVI_p_col_zoo
for (i in c(1:length(R_Active_GARCH11SVI_p[,P_number])))
{R_Active_GARCH11SVI_p_cumulated[i,P_number] = prod((1+R_Active_GARCH11SVI_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_GARCH11SVI_p_col_zoo*0)))[1:i])}
}
GARCH11SVI_CandD_3D_HighRes = (plot_ly(z=R_Active_GARCH11SVI_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using GARCH11SVI models where 0.4<=P<=0.55",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(GARCH11SVI_CandD_3D, filename = "GARCH11SVI_CandD_3D-public-graph")
###----------------------------------------------------------------------------
### FTSE Granger and Pesaran (2000)'s framework using the GJRGARCH11 model (p=0.505 seems best)
###----------------------------------------------------------------------------
p = 0.505
# corresponding directional forecast and realised direction
y_hat_d_GJRGARCH11 = ifelse(GJRGARCH11_CandD_zoo_no_nas>p,1,0)
y_d_GJRGARCH11 = y_d - (GJRGARCH11_CandD_zoo_no_nas*0) # This y_d only has
# values on dates corresponding to GJRGARCH11's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_GJRGARCH11),1))==1 , 1, 0)
R_Active_GJRGARCH11_p = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_GJRGARCH11*0))+(BoE_r_f_zoo-(y_hat_d_GJRGARCH11*0))) +
(1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_GJRGARCH11*0))))[2:length(y_hat_d_GJRGARCH11)]
# note that the trick '(FTSE_R_zoo-(y_hat_d_GJRGARCH11*0))' returns only the R values
# corresponding to dates present in y_hat_d_GARCH11, which are the ones
# in omega (omega couldn't be used because it's not a zoo object)
R_Active_GJRGARCH11_p_cumulated = matrix(,nrow=length(R_Active_GJRGARCH11_p))
for (i in c(1:length(R_Active_GJRGARCH11_p)))
{R_Active_GJRGARCH11_p_cumulated[i] = prod((1+R_Active_GJRGARCH11_p)[1:i])}
# plot(R_Active_GJRGARCH11_p_cumulated, type='l')
R_cumulated = matrix(,nrow=length(R_Active_GJRGARCH11_p))
for (i in c(1:length(R_Active_GJRGARCH11_p)))
{R_cumulated[i] = prod((1+(FTSE_R_zoo+(BoE_r_f_zoo - (R_Active_GJRGARCH11_p*0))))[1:i])}
# plot(R_cumulated, type='l')
plot(zoo(R_Active_GJRGARCH11_p_cumulated,
as.Date(zoo::index(y_hat_d_GJRGARCH11[2:length(y_hat_d_GJRGARCH11)]))),
type="l",col="red", xlab='Date', main=p,
ylab='FTSE index and GJRGARCH11 index cumulated')
lines(zoo(R_cumulated,
as.Date(zoo::index(y_hat_d_GJRGARCH11[2:length(y_hat_d_GJRGARCH11)]))),
col="black")
lines(zoo(R_Active_Naive_p_cumulated,
as.Date(zoo::index(y_hat_d_Naive[2:length(y_hat_d_Naive)]))),
col="green")
##----------------------------------------------------------------------------
## 3D graph Low Resolution
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
sequance = seq(from = 0, to = 1, by = 0.05)
R_Active_GJRGARCH11_p = matrix(, nrow=3585, ncol=length(sequance))
R_Active_GJRGARCH11_p_cumulated = matrix(, nrow=3585, ncol=length(sequance))
R_cumulated = matrix(, nrow=3585, ncol=length(sequance))
y_d_GJRGARCH11 = y_d - (GJRGARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GJRGARCH11's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(length(sequance)-1)
# corresponding directional forecast and realised direction
y_hat_d_GJRGARCH11 = ifelse(GJRGARCH11_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_GJRGARCH11),1))==1 , 1, 0)
R_Active_GJRGARCH11_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_GJRGARCH11*0))+(BoE_r_f_zoo-(y_hat_d_GJRGARCH11*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_GJRGARCH11*0))))[2:length(y_hat_d_GJRGARCH11)] # note that the trick '(FTSE_R_zoo-(y_hat_d_GJRGARCH11*0))' returns only the R values corresponding to dates present in y_hat_d_GJRGARCH11, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_GJRGARCH11_p[,P_number] = R_Active_GJRGARCH11_p_col_zoo
for (i in c(1:length(R_Active_GJRGARCH11_p[,P_number])))
{R_Active_GJRGARCH11_p_cumulated[i,P_number] = prod((1+R_Active_GJRGARCH11_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_GJRGARCH11_p_col_zoo*0)))[1:i])}
}
GJRGARCH11_CandD_3D_LowRes = (plot_ly(z=R_Active_GJRGARCH11_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using GJRGARCH11 models",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(GJRGARCH11_CandD_3D, filename = "GJRGARCH11_CandD_3D-public-graph")
##----------------------------------------------------------------------------
## 3D graph High Resolution
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
BY = 0.0125
FROM = 0.45
sequance = seq(from = FROM, to = 0.6, by = BY)
R_Active_GJRGARCH11_p = matrix(, nrow=3585, ncol=length(sequance))
R_Active_GJRGARCH11_p_cumulated = matrix(, nrow=3585, ncol=length(sequance))
R_cumulated = matrix(, nrow=3585, ncol=length(sequance))
y_d_GJRGARCH11 = y_d - (GJRGARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GJRGARCH11's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(1/BY) - FROM/BY
# corresponding directional forecast and realised direction
y_hat_d_GJRGARCH11 = ifelse(GJRGARCH11_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_GJRGARCH11),1))==1 , 1, 0)
R_Active_GJRGARCH11_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_GJRGARCH11*0))+ (BoE_r_f_zoo-(y_hat_d_GJRGARCH11*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_GJRGARCH11*0))))[2:length(y_hat_d_GJRGARCH11)] # note that the trick '(FTSE_R_zoo-(y_hat_d_GJRGARCH11*0))' returns only the R values corresponding to dates present in y_hat_d_GJRGARCH11, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_GJRGARCH11_p[,P_number] = R_Active_GJRGARCH11_p_col_zoo
for (i in c(1:length(R_Active_GJRGARCH11_p[,P_number])))
{R_Active_GJRGARCH11_p_cumulated[i,P_number] = prod((1+R_Active_GJRGARCH11_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_GJRGARCH11_p_col_zoo*0)))[1:i])}
}
GJRGARCH11_CandD_3D_HighRes = (plot_ly(z=R_Active_GJRGARCH11_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using GJRGARCH11 models where 0.45<=P<=0.6",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(GJRGARCH11_CandD_3D, filename = "GJRGARCH11_CandD_3D-public-graph")
###----------------------------------------------------------------------------
### FTSE Granger and Pesaran (2000)'s framework using the GJRGARCH11SVI model (p=0.506 seems best)
###----------------------------------------------------------------------------
p = 0.506
# corresponding directional forecast and realised direction
y_hat_d_GJRGARCH11SVI = ifelse(GJRGARCH11SVI_CandD_zoo_no_nas>p,1,0)
y_d_GJRGARCH11SVI = y_d - (GJRGARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has
# values on dates corresponding to GJRGARCH11SVI's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_GJRGARCH11SVI),1))==1 , 1, 0)
R_Active_GJRGARCH11SVI_p = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_GJRGARCH11SVI*0))+(BoE_r_f_zoo-(y_hat_d_GJRGARCH11SVI*0))) +
(1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_GJRGARCH11SVI*0))))[2:length(y_hat_d_GJRGARCH11SVI)]
# note that the trick '(FTSE_R_zoo-(y_hat_d_GJRGARCH11SVI*0))' returns only the R values
# corresponding to dates present in y_hat_d_GARCH11, which are the ones
# in omega (omega couldn't be used because it's not a zoo object)
R_Active_GJRGARCH11SVI_p_cumulated = matrix(,nrow=length(R_Active_GJRGARCH11SVI_p))
for (i in c(1:length(R_Active_GJRGARCH11SVI_p)))
{R_Active_GJRGARCH11SVI_p_cumulated[i] = prod((1+R_Active_GJRGARCH11SVI_p)[1:i])}
# plot(R_Active_GJRGARCH11SVI_p_cumulated, type='l')
R_cumulated = matrix(,nrow=length(R_Active_GJRGARCH11SVI_p))
for (i in c(1:length(R_Active_GJRGARCH11SVI_p)))
{R_cumulated[i] = prod((1+(FTSE_R_zoo+(BoE_r_f_zoo - (R_Active_GJRGARCH11SVI_p*0))))[1:i])}
# plot(R_cumulated, type='l')
plot(zoo(R_Active_GJRGARCH11SVI_p_cumulated,
as.Date(zoo::index(y_hat_d_GJRGARCH11SVI[2:length(y_hat_d_GJRGARCH11SVI)]))),
type="l",col="red", xlab='Date', main = p,
ylab='FTSE index and GJRGARCH11SVI index cumulated')
lines(zoo(R_cumulated,
as.Date(zoo::index(y_hat_d_GJRGARCH11SVI[2:length(y_hat_d_GJRGARCH11SVI)]))),
col="black")
lines(zoo(R_Active_Naive_p_cumulated,
as.Date(zoo::index(y_hat_d_Naive[2:length(y_hat_d_Naive)]))),
col="green")
##----------------------------------------------------------------------------
## 3D graph Low Resolution
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
sequance = seq(from = 0, to = 1, by = 0.05)
R_Active_GJRGARCH11SVI_p = matrix(, nrow=3567, ncol=length(sequance))
R_Active_GJRGARCH11SVI_p_cumulated = matrix(, nrow=3567, ncol=length(sequance))
R_cumulated = matrix(, nrow=3567, ncol=length(sequance))
y_d_GJRGARCH11SVI = y_d - (GJRGARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GJRGARCH11SVI's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(length(sequance)-1)
# corresponding directional forecast and realised direction
y_hat_d_GJRGARCH11SVI = ifelse(GJRGARCH11SVI_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_GJRGARCH11SVI),1))==1 , 1, 0)
R_Active_GJRGARCH11SVI_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_GJRGARCH11SVI*0))+(BoE_r_f_zoo-(y_hat_d_GJRGARCH11SVI*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_GJRGARCH11SVI*0))))[2:length(y_hat_d_GJRGARCH11SVI)] # note that the trick '(FTSE_R_zoo-(y_hat_d_GJRGARCH11SVI*0))' returns only the R values corresponding to dates present in y_hat_d_GJRGARCH11SVI, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_GJRGARCH11SVI_p[,P_number] = R_Active_GJRGARCH11SVI_p_col_zoo
for (i in c(1:length(R_Active_GJRGARCH11SVI_p[,P_number])))
{R_Active_GJRGARCH11SVI_p_cumulated[i,P_number] = prod((1+R_Active_GJRGARCH11SVI_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_GJRGARCH11SVI_p_col_zoo*0)))[1:i])}
}
GJRGARCH11SVI_CandD_3D_LowRes = (plot_ly(z=R_Active_GJRGARCH11SVI_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using GJRGARCH11SVI models",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(GJRGARCH11SVI_CandD_3D, filename = "GJRGARCH11SVI_CandD_3D-public-graph")
##----------------------------------------------------------------------------
## 3D graph High Resolution
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
BY = 0.0125
FROM = 0.45
sequance = seq(from = FROM, to = 0.6, by = BY)
R_Active_GJRGARCH11SVI_p = matrix(, nrow=3567, ncol=length(sequance))
R_Active_GJRGARCH11SVI_p_cumulated = matrix(, nrow=3567, ncol=length(sequance))
R_cumulated = matrix(, nrow=3567, ncol=length(sequance))
y_d_GJRGARCH11SVI = y_d - (GJRGARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GJRGARCH11SVI's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(1/BY) - FROM/BY
# corresponding directional forecast and realised direction
y_hat_d_GJRGARCH11SVI = ifelse(GJRGARCH11SVI_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_GJRGARCH11SVI),1))==1 , 1, 0)
R_Active_GJRGARCH11SVI_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_GJRGARCH11SVI*0))+ (BoE_r_f_zoo-(y_hat_d_GJRGARCH11SVI*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_GJRGARCH11SVI*0))))[2:length(y_hat_d_GJRGARCH11SVI)] # note that the trick '(FTSE_R_zoo-(y_hat_d_GJRGARCH11SVI*0))' returns only the R values corresponding to dates present in y_hat_d_GJRGARCH11SVI, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_GJRGARCH11SVI_p[,P_number] = R_Active_GJRGARCH11SVI_p_col_zoo
for (i in c(1:length(R_Active_GJRGARCH11SVI_p[,P_number])))
{R_Active_GJRGARCH11SVI_p_cumulated[i,P_number] = prod((1+R_Active_GJRGARCH11SVI_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_GJRGARCH11SVI_p_col_zoo*0)))[1:i])}
}
GJRGARCH11SVI_CandD_3D_HighRes = (plot_ly(z=R_Active_GJRGARCH11SVI_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using GJRGARCH11SVI models where 0.45<=P<=0.6",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(GJRGARCH11SVI_CandD_3D, filename = "GJRGARCH11SVI_CandD_3D-public-graph")
###----------------------------------------------------------------------------
### FTSE Granger and Pesaran (2000)'s framework using the EGARCH11 model (p=0.499 seems best)
###----------------------------------------------------------------------------
p=0.499
# corresponding directional forecast and realised direction
y_hat_d_EGARCH11 = ifelse(EGARCH11_CandD_zoo_no_nas>p,1,0)
y_d_EGARCH11 = y_d - (EGARCH11_CandD_zoo_no_nas*0) # This y_d only has
# values on dates corresponding to EGARCH11's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_EGARCH11),1))==1 , 1, 0)
R_Active_EGARCH11_p = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_EGARCH11*0))+(BoE_r_f_zoo-(y_hat_d_EGARCH11*0))) +
(1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_EGARCH11*0))))[2:length(y_hat_d_EGARCH11)]
# note that the trick '(FTSE_R_zoo-(y_hat_d_EGARCH11*0))' returns only the R values
# corresponding to dates present in y_hat_d_EGARCH11, which are the ones
# in omega (omega couldn't be used because it's not a zoo object)
R_Active_EGARCH11_p_cumulated = matrix(,nrow=length(R_Active_EGARCH11_p))
for (i in c(1:length(R_Active_EGARCH11_p)))
{R_Active_EGARCH11_p_cumulated[i] = prod((1+R_Active_EGARCH11_p)[1:i])}
# plot(R_Active_EGARCH11_p_cumulated, type='l')
R_cumulated = matrix(,nrow=length(R_Active_EGARCH11_p))
for (i in c(1:length(R_Active_EGARCH11_p)))
{R_cumulated[i] = prod((1+(FTSE_R_zoo+(BoE_r_f_zoo - (R_Active_EGARCH11_p*0))))[1:i])}
# plot(R_cumulated, type='l')
plot(zoo(R_Active_EGARCH11_p_cumulated,
as.Date(zoo::index(y_hat_d_EGARCH11[2:length(y_hat_d_EGARCH11)]))),
type="l",col="red", xlab='Date', main = p,
ylab='FTSE index and EGARCH11 index cumulated')
lines(zoo(R_cumulated,
as.Date(zoo::index(y_hat_d_EGARCH11[2:length(y_hat_d_EGARCH11)]))),
col="black")
lines(zoo(R_Active_Naive_p_cumulated,
as.Date(zoo::index(y_hat_d_Naive[2:length(y_hat_d_Naive)]))),
col="green")
##----------------------------------------------------------------------------
## 3D graph Low Resolution
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
sequance = seq(from = 0, to = 1, by = 0.05)
R_Active_EGARCH11_p = matrix(, nrow=3585, ncol=length(sequance))
R_Active_EGARCH11_p_cumulated = matrix(, nrow=3585, ncol=length(sequance))
R_cumulated = matrix(, nrow=3585, ncol=length(sequance))
y_d_EGARCH11 = y_d - (EGARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to EGARCH11's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(length(sequance)-1)
# corresponding directional forecast and realised direction
y_hat_d_EGARCH11 = ifelse(EGARCH11_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_EGARCH11),1))==1 , 1, 0)
R_Active_EGARCH11_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_EGARCH11*0))+(BoE_r_f_zoo-(y_hat_d_EGARCH11*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_EGARCH11*0))))[2:length(y_hat_d_EGARCH11)] # note that the trick '(FTSE_R_zoo-(y_hat_d_EGARCH11*0))' returns only the R values corresponding to dates present in y_hat_d_EGARCH11, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_EGARCH11_p[,P_number] = R_Active_EGARCH11_p_col_zoo
for (i in c(1:length(R_Active_EGARCH11_p[,P_number])))
{R_Active_EGARCH11_p_cumulated[i,P_number] = prod((1+R_Active_EGARCH11_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_EGARCH11_p_col_zoo*0)))[1:i])}
}
EGARCH11_CandD_3D_LowRes = (plot_ly(z=R_Active_EGARCH11_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using EGARCH11 models",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(EGARCH11_CandD_3D, filename = "EGARCH11_CandD_3D-public-graph")
##----------------------------------------------------------------------------
## 3D graph High Resolution
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
BY = 0.0125
FROM = 0.45
sequance = seq(from = FROM, to = 0.6, by = BY)
R_Active_EGARCH11_p = matrix(, nrow=3585, ncol=length(sequance))
R_Active_EGARCH11_p_cumulated = matrix(, nrow=3585, ncol=length(sequance))
R_cumulated = matrix(, nrow=3585, ncol=length(sequance))
y_d_EGARCH11 = y_d - (EGARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to EGARCH11's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(1/BY) - FROM/BY
# corresponding directional forecast and realised direction
y_hat_d_EGARCH11 = ifelse(EGARCH11_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_EGARCH11),1))==1 , 1, 0)
R_Active_EGARCH11_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_EGARCH11*0))+ (BoE_r_f_zoo-(y_hat_d_EGARCH11*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_EGARCH11*0))))[2:length(y_hat_d_EGARCH11)] # note that the trick '(FTSE_R_zoo-(y_hat_d_EGARCH11*0))' returns only the R values corresponding to dates present in y_hat_d_EGARCH11, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_EGARCH11_p[,P_number] = R_Active_EGARCH11_p_col_zoo
for (i in c(1:length(R_Active_EGARCH11_p[,P_number])))
{R_Active_EGARCH11_p_cumulated[i,P_number] = prod((1+R_Active_EGARCH11_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_EGARCH11_p_col_zoo*0)))[1:i])}
}
EGARCH11_CandD_3D_HighRes = (plot_ly(z=R_Active_EGARCH11_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using an EGARCH11 model where 0.45<=P<=0.6",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(EGARCH11_CandD_3D, filename = "EGARCH11_CandD_3D-public-graph")
###----------------------------------------------------------------------------
### FTSE Granger and Pesaran (2000)'s framework using the EGARCH11-SVI model (p=0.46 seems best)
###----------------------------------------------------------------------------
p=0.46
# corresponding directional forecast and realised direction
y_hat_d_EGARCH11SVI = ifelse(EGARCH11SVI_CandD_zoo_no_nas>p,1,0)
y_d_EGARCH11SVI = y_d - (EGARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has
# values on dates corresponding to EGARCH11SVI's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_EGARCH11SVI),1))==1 , 1, 0)
R_Active_p = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_EGARCH11SVI*0))+
(BoE_r_f_zoo-(y_hat_d_EGARCH11SVI*0))) +
(1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_EGARCH11SVI*0))))[2:length(y_hat_d_EGARCH11SVI)]
# note that the trick '(FTSE_R_zoo-(y_hat_d_EGARCH11SVI*0))' returns only the R values
# corresponding to dates present in y_hat_d_EGARCH11SVI, which are the ones
# in omega (omega couldn't be used because it's not a zoo object)
R_Active_p_cumulated = matrix(,nrow=length(R_Active_p))
for (i in c(1:length(R_Active_p)))
{R_Active_p_cumulated[i] = prod((1+R_Active_p)[1:i])}
# plot(R_Active_p_cumulated, type='l')
R_cumulated = matrix(,nrow=length(R_Active_p))
for (i in c(1:length(R_Active_p)))
{R_cumulated[i] = prod((1+(FTSE_R_zoo+(BoE_r_f_zoo - (R_Active_p*0))))[1:i])}
# plot(R_cumulated, type='l')
plot(zoo(R_Active_p_cumulated,
as.Date(zoo::index(y_hat_d_EGARCH11SVI[2:length(y_hat_d_EGARCH11SVI)]))),
type="l",col="red", xlab='Date', main = p,
ylab='FTSE index and EGARCH11SVI index cumulated')
lines(zoo(R_cumulated,
as.Date(zoo::index(y_hat_d_EGARCH11SVI[2:length(y_hat_d_EGARCH11SVI)]))),
col="black")
lines(zoo(R_Active_Naive_p_cumulated,
as.Date(zoo::index(y_hat_d_Naive[2:length(y_hat_d_Naive)]))),
col="green")
##----------------------------------------------------------------------------
## 3D graph Low Resolution
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
sequance = seq(from = 0, to = 1, by = 0.05)
R_Active_EGARCH11SVI_p = matrix(, nrow=3565, ncol=length(sequance))
R_Active_EGARCH11SVI_p_cumulated = matrix(, nrow=3565, ncol=length(sequance))
R_cumulated = matrix(, nrow=3565, ncol=length(sequance))
y_d_EGARCH11SVI = y_d - (EGARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to EGARCH11SVI's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(length(sequance)-1)
# corresponding directional forecast and realised direction
y_hat_d_EGARCH11SVI = ifelse(EGARCH11SVI_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_EGARCH11SVI),1))==1 , 1, 0)
R_Active_EGARCH11SVI_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_EGARCH11SVI*0))+(BoE_r_f_zoo-(y_hat_d_EGARCH11SVI*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_EGARCH11SVI*0))))[2:length(y_hat_d_EGARCH11SVI)] # note that the trick '(FTSE_R_zoo-(y_hat_d_EGARCH11SVI*0))' returns only the R values corresponding to dates present in y_hat_d_EGARCH11SVI, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_EGARCH11SVI_p[,P_number] = R_Active_EGARCH11SVI_p_col_zoo
for (i in c(1:length(R_Active_EGARCH11SVI_p[,P_number])))
{R_Active_EGARCH11SVI_p_cumulated[i,P_number] = prod((1+R_Active_EGARCH11SVI_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_EGARCH11SVI_p_col_zoo*0)))[1:i])}
}
EGARCH11SVI_CandD_3D_LowRes = (plot_ly(z=R_Active_EGARCH11SVI_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using EGARCH11SVI models",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(EGARCH11SVI_CandD_3D, filename = "EGARCH11SVI_CandD_3D-public-graph")
##----------------------------------------------------------------------------
## 3D graph High Resolution
##----------------------------------------------------------------------------
# Set the probability above which we would like our investing algorythm to invest in our 3d grath
BY = 0.0125
FROM = 0.45
sequance = seq(from = FROM, to = 0.6, by = BY)
R_Active_EGARCH11SVI_p = matrix(, nrow=3565, ncol=length(sequance))
R_Active_EGARCH11SVI_p_cumulated = matrix(, nrow=3565, ncol=length(sequance))
R_cumulated = matrix(, nrow=3565, ncol=length(sequance))
y_d_EGARCH11SVI = y_d - (EGARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to EGARCH11SVI's.
# Set eh probability above which we would like our investing algorythm to invest
for(P in sequance)
{P_number = 1 + P*(1/BY) - FROM/BY
# corresponding directional forecast and realised direction
y_hat_d_EGARCH11SVI = ifelse(EGARCH11SVI_CandD_zoo_no_nas>P,1,0)
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_EGARCH11SVI),1))==1 , 1, 0)
R_Active_EGARCH11SVI_p_col_zoo = ((lag(omega,1) * ((FTSE_R_zoo-(y_hat_d_EGARCH11SVI*0))+ (BoE_r_f_zoo-(y_hat_d_EGARCH11SVI*0))) + (1-lag(omega,1)) * (BoE_r_f_zoo-(y_hat_d_EGARCH11SVI*0))))[2:length(y_hat_d_EGARCH11SVI)] # note that the trick '(FTSE_R_zoo-(y_hat_d_EGARCH11SVI*0))' returns only the R values corresponding to dates present in y_hat_d_EGARCH11SVI, which are the ones in omega (omega couldn't be used because it's not a zoo object)
R_Active_EGARCH11SVI_p[,P_number] = R_Active_EGARCH11SVI_p_col_zoo
for (i in c(1:length(R_Active_EGARCH11SVI_p[,P_number])))
{R_Active_EGARCH11SVI_p_cumulated[i,P_number] = prod((1+R_Active_EGARCH11SVI_p[,P_number])[1:i])
R_cumulated[i,P_number] = prod((1+(FTSE_R_zoo-(R_Active_EGARCH11SVI_p_col_zoo*0)))[1:i])}
}
EGARCH11SVI_CandD_3D_HighRes = (plot_ly(z=R_Active_EGARCH11SVI_p_cumulated, type="scatter3d") %>%
layout(
title = "Portfolio Investing £1 as per the Active C&D Model Using EGARCH11SVI models where 0.45<=P<=0.6",
scene = list(
xaxis = list(title = "Investment Threashold"),
yaxis = list(title = "Date"),
zaxis = list(title = "Portfolio Return")
)) %>%
add_surface())
# chart_link = api_create(EGARCH11SVI_CandD_3D, filename = "EGARCH11SVI_CandD_3D-public-graph")
###----------------------------------------------------------------------------
### FTSE Granger and Pesaran (2000)'s framework 2d Graphs put together
###----------------------------------------------------------------------------
# Plot all 2d polts
plot(zoo(R_Active_GARCH11_p_cumulated_2d,
as.Date(zoo::index(y_hat_d_GARCH11_2d[2:length(y_hat_d_GARCH11_2d)]))),
cex.axis=1, type="l",col="red", xlab='Date', ylim=c(0.67,2.05), ylab='£')
lines(zoo(R_cumulated_2d,
as.Date(zoo::index(y_hat_d_GARCH11_2d[2:length(y_hat_d_GARCH11_2d)]))),
col="black")
lines(zoo(R_Active_GARCH11SVI_p_cumulated, as.Date(zoo::index(y_hat_d_GARCH11SVI[2:length(y_hat_d_GARCH11SVI)]))),
col="orange")
lines(zoo(R_Active_GJRGARCH11_p_cumulated,
as.Date(zoo::index(y_hat_d_GJRGARCH11[2:length(y_hat_d_GJRGARCH11)]))),
col="blue")
lines(zoo(R_Active_GJRGARCH11SVI_p_cumulated,
as.Date(zoo::index(y_hat_d_GJRGARCH11SVI[2:length(y_hat_d_GJRGARCH11SVI)]))),
col="magenta")
lines(zoo(R_Active_EGARCH11_p_cumulated,
as.Date(zoo::index(y_hat_d_EGARCH11[2:length(y_hat_d_EGARCH11)]))),
col="green")
lines(zoo(R_Active_p_cumulated,
as.Date(zoo::index(y_hat_d_EGARCH11SVI[2:length(y_hat_d_EGARCH11SVI)]))),
col="purple")
lines(zoo(R_Active_Naive_p_cumulated,
as.Date(zoo::index(y_hat_d_Naive[2:length(y_hat_d_Naive)]))),
col="bisque4")
legend(lty=1, cex=1,
"topleft", col=c("black", "bisque4", "red", "orange", "blue", "magenta", "green", "purple"),
legend=c("Buy and hold", "Naïve", "GARCH for psi=0.4900", "GARCH-SVI for psi=0.4935", "GJRGARCH for psi=0.5050", "GJRGARCH-SVI for psi=0.5060", "EGARCH for psi=0.4990", "EGARCH-SVI for psi=0.4600"))
####---------------------------------------------------------------------------
#### SPX1 GARCH models: create, train/fit them and create forecasts
####---------------------------------------------------------------------------
# Set parameters
SPX1_roll = length(SPX_dSVI_zoo)-247
# This will also be the number of out-of-sample predictions. N.B.: the 248th
# value of R corresponds to 2005-01-03, the 1st trading day out-of-sample,
# the first value after 247.
###----------------------------------------------------------------------------
### SPX1 In-sample SPX1_AR(1)-GARCH(1,1) Model
###---------------------------------------------------------------------------
SPX1_in_sample_GARCH11 = GARCH_model_spec(mod="sGARCH", exreg= NULL)
SPX1_in_sample_GARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPX1_in_sample_GARCH11)
###----------------------------------------------------------------------------
### Create the SPX1_AR(1)-GARCH(1,1) Model and forecasts
###----------------------------------------------------------------------------
SPX1_GARCH11_mu_hat = zoo(matrix(, nrow=roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(SPX1_GARCH11_mu_hat) = c('SPX1_GARCH11_mu_hat')
SPX1_GARCH11_sigma_hat_test = zoo(matrix(, nrow=roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(GARCH11_sigma_hat) = c('GARCH11_sigma_hat')
for (i in c(1:SPX1_roll))
{SPX1_GARCH11 = GARCH_model_spec(mod="sGARCH", exreg=NULL)
try(withTimeout({(SPX1_GARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPX1_GARCH11))}, timeout = 5),silent = TRUE)
# the 247th value of R is 2004-12-31.
try((SPX1_GARCH11fitforecast = GARCH_model_forecast(mod=SPX1_GARCH11fit)), silent = TRUE)
# suppressWarnings(expr)
try((SPX1_GARCH11_mu_hat[i]=SPX1_GARCH11fitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPX1_GARCH11_sigma_hat[i]=SPX1_GARCH11fitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPX1_GARCH11)
rm(SPX1_GARCH11fit)
rm(SPX1_GARCH11fitforecast)
}
# fitted(GARCH11fitforecast)
SPX_RV_no_SPX1_GARCH11_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPX1_roll)], as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))
-(na.omit(zoo((SPX1_GARCH11_sigma_hat*0),
as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))))))
# This above is just a neat trick that returns strictly the RV values corresponding to the dates for which GARCH11_sigma_hat values are not NA without any new functions or packages.
# Forecast Error, Mean and S.D.:
SPX1_GARCH11forecasterror = (na.omit(SPX1_GARCH11_sigma_hat)
-SPX_RV_no_SPX1_GARCH11_sigma_hat_na_dated)
colnames(SPX1_GARCH11forecasterror) = c("SPX1_GARCH11forecasterror")
# plot(zoo(GARCH11forecasterror, as.Date(row.names(GARCH11forecasterror))), type='l', ylab='GARCH11forecasterror', xlab='Date')
mean(SPX1_GARCH11forecasterror)
sd(SPX1_GARCH11forecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPX1_GARCH11_sigma_hat_na_dated, (na.omit(SPX1_GARCH11_sigma_hat)))
###----------------------------------------------------------------------------
### SPX1 In-sample SPX1_AR(1)-GARCH(1,1)-SVI Model
###---------------------------------------------------------------------------
SPX1_in_sample_GARCH11SVI = GARCH_model_spec(mod="sGARCH", exreg= as.matrix(SPX_dSVI_zoo[1:(246+1)]))
SPX1_in_sample_GARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPX1_in_sample_GARCH11SVI)
###----------------------------------------------------------------------------
### Create the SPX1_AR(1)-GARCH(1,1)-SVI Model and forecasts
###----------------------------------------------------------------------------
SPX1_GARCH11SVI_mu_hat = zoo(matrix(, nrow=SPX1_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(SPX1_GARCH11SVI_mu_hat) = c('SPX1_GARCH11_mu_hat')
SPX1_GARCH11SVI_sigma_hat = zoo(matrix(, nrow=SPX1_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(GARCH11_sigma_hat) = c('GARCH11_sigma_hat')
for (i in c(1:SPX1_roll))
{SPX1_GARCH11SVI = GARCH_model_spec(mod="sGARCH", exreg=as.matrix(SPX_dSVI_zoo[1:(246+i)]))
try(withTimeout({(SPX1_GARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPX1_GARCH11SVI))}, timeout = 5),silent = TRUE)
try((SPX1_GARCH11SVIfitforecast = GARCH_model_forecast(mod=SPX1_GARCH11SVIfit)), silent = TRUE)
try((SPX1_GARCH11SVI_mu_hat[i]=SPX1_GARCH11SVIfitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPX1_GARCH11SVI_sigma_hat[i]=SPX1_GARCH11SVIfitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPX1_GARCH11SVI)
rm(SPX1_GARCH11SVIfit)
rm(SPX1_GARCH11SVIfitforecast)
}
# fitted(GARCH11SVIfitforecast)
SPX_RV_no_SPX1_GARCH11SVI_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPX1_roll)], as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))
-(na.omit(zoo((SPX1_GARCH11SVI_sigma_hat*0), as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))))))
# Forecast Error, Mean and S.D.:
SPX1_GARCH11SVIforecasterror = (na.omit(SPX1_GARCH11SVI_sigma_hat)
-SPX_RV_no_SPX1_GARCH11SVI_sigma_hat_na_dated)
colnames(SPX1_GARCH11SVIforecasterror) = c("SPX1_GARCH11SVIforecasterror")
# plot(zoo(GARCH11SVIforecasterror, as.Date(row.names(GARCH11SVIforecasterror))), type='l', ylab='GARCH11SVIforecasterror', xlab='Date')
mean(SPX1_GARCH11SVIforecasterror)
sd(SPX1_GARCH11SVIforecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPX1_GARCH11SVI_sigma_hat_na_dated, (na.omit(SPX1_GARCH11SVI_sigma_hat)))
# Note that this is the same as the bellow:
# sqrt(mean((RV[(249+1):(length(R_vector)+1)] - as.matrix((GARCH11SVIfitforecast@forecast[["sigmaFor"]]))[1:SPX1_roll])^2))
###----------------------------------------------------------------------------
### SPX1 In-sample SPX1_AR(1)-GJRGARCH(1,1) Model
###---------------------------------------------------------------------------
SPX1_in_sample_GJRGARCH11 = GARCH_model_spec(mod="gjrGARCH", exreg= NULL)
SPX1_in_sample_GJRGARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPX1_in_sample_GJRGARCH11)
###----------------------------------------------------------------------------
### Create the SPX1_AR(1)-GJRGARCH(1,1) Model and forecasts
###----------------------------------------------------------------------------
SPX1_GJRGARCH11_mu_hat = zoo(matrix(, nrow=SPX1_roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(SPX1_GJRGARCH11_mu_hat) = c('SPX1_GJRGARCH11_mu_hat')
SPX1_GJRGARCH11_sigma_hat = zoo(matrix(, nrow=SPX1_roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(GJRGARCH11_sigma_hat) = c('GJRGARCH11_sigma_hat')
for (i in c(1:SPX1_roll))
{SPX1_GJRGARCH11 = GARCH_model_spec(mod="gjrGARCH", exreg=NULL)
try(withTimeout({(SPX1_GJRGARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPX1_GJRGARCH11))}, timeout = 5),silent = TRUE)
# the 247th value of R is 2004-12-31.
try((SPX1_GJRGARCH11fitforecast = GARCH_model_forecast(mod=SPX1_GJRGARCH11fit)), silent = TRUE)
# suppressWarnings(expr)
try((SPX1_GJRGARCH11_mu_hat[i]=SPX1_GJRGARCH11fitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPX1_GJRGARCH11_sigma_hat[i]=SPX1_GJRGARCH11fitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPX1_GJRGARCH11)
rm(SPX1_GJRGARCH11fit)
rm(SPX1_GJRGARCH11fitforecast)
}
# fitted(GJRGARCH11fitforecast)
SPX_RV_no_SPX1_GJRGARCH11_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPX1_roll)], as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))
-(na.omit(zoo((SPX1_GJRGARCH11_sigma_hat*0),
as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))))))
# Forecast Error, Mean and S.D.:
SPX1_GJRGARCH11forecasterror = (na.omit(SPX1_GJRGARCH11_sigma_hat)
-SPX_RV_no_SPX1_GJRGARCH11_sigma_hat_na_dated)
colnames(SPX1_GJRGARCH11forecasterror) = c("SPX1_GJRGARCH11forecasterror")
# plot(zoo(GJRGARCH11forecasterror, as.Date(row.names(GJRGARCH11forecasterror))), type='l', ylab='GJRGARCH11forecasterror', xlab='Date')
mean(SPX1_GJRGARCH11forecasterror)
sd(SPX1_GJRGARCH11forecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPX1_GJRGARCH11_sigma_hat_na_dated, (na.omit(SPX1_GJRGARCH11_sigma_hat)))
###----------------------------------------------------------------------------
### SPX1 In-sample SPX1_AR(1)-GJRGARCH(1,1)-SVI Model
###---------------------------------------------------------------------------
SPX1_in_sample_GJRGARCH11SVI = GARCH_model_spec(mod="gjrGARCH", exreg= as.matrix(SPX_dSVI_zoo[1:(246+1)]))
SPX1_in_sample_GJRGARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPX1_in_sample_GJRGARCH11SVI)
###----------------------------------------------------------------------------
### Create the SPX1_AR(1)-GJRGARCH(1,1)-SVI Model and forecasts
###----------------------------------------------------------------------------
SPX1_GJRGARCH11SVI_mu_hat = zoo(matrix(, nrow=SPX1_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(SPX1_GJRGARCH11SVI_mu_hat) = c('SPX1_GJRGARCH11_mu_hat')
SPX1_GJRGARCH11SVI_sigma_hat = zoo(matrix(, nrow=SPX1_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(GJRGARCH11_sigma_hat) = c('GJRGARCH11_sigma_hat')
for (i in c(1:SPX1_roll))
{SPX1_GJRGARCH11SVI = GARCH_model_spec(mod="gjrGARCH", exreg=as.matrix(SPX_dSVI_zoo[1:(246+i)]))
try(withTimeout({(SPX1_GJRGARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPX1_GJRGARCH11SVI))}, timeout = 5),silent = TRUE)
try((SPX1_GJRGARCH11SVIfitforecast = GARCH_model_forecast(mod=SPX1_GJRGARCH11SVIfit)), silent = TRUE)
try((SPX1_GJRGARCH11SVI_mu_hat[i]=SPX1_GJRGARCH11SVIfitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPX1_GJRGARCH11SVI_sigma_hat[i]=SPX1_GJRGARCH11SVIfitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPX1_GJRGARCH11SVI)
rm(SPX1_GJRGARCH11SVIfit)
rm(SPX1_GJRGARCH11SVIfitforecast)
}
# fitted(GJRGARCH11SVIfitforecast)
SPX_RV_no_SPX1_GJRGARCH11SVI_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPX1_roll)], as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))
-(na.omit(zoo((SPX1_GJRGARCH11SVI_sigma_hat*0), as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))))))
# Forecast Error, Mean and S.D.:
SPX1_GJRGARCH11SVIforecasterror = (na.omit(SPX1_GJRGARCH11SVI_sigma_hat)
-SPX_RV_no_SPX1_GJRGARCH11SVI_sigma_hat_na_dated)
colnames(SPX1_GJRGARCH11SVIforecasterror) = c("SPX1_GJRGARCH11SVIforecasterror")
# plot(zoo(GJRGARCH11SVIforecasterror, as.Date(row.names(GJRGARCH11SVIforecasterror))), type='l', ylab='GJRGARCH11SVIforecasterror', xlab='Date')
mean(SPX1_GJRGARCH11SVIforecasterror)
sd(SPX1_GJRGARCH11SVIforecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPX1_GJRGARCH11SVI_sigma_hat_na_dated, (na.omit(SPX1_GJRGARCH11SVI_sigma_hat)))
###----------------------------------------------------------------------------
### SPX1 In-sample SPX1_AR(1)-EGARCH(1,1) Model
###---------------------------------------------------------------------------
SPX1_in_sample_EGARCH11 = GARCH_model_spec(mod="eGARCH", exreg= NULL)
SPX1_in_sample_EGARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPX1_in_sample_EGARCH11)
###----------------------------------------------------------------------------
### Create the SPX1_AR(1)-EGARCH(1,1) Model and forecasts
###----------------------------------------------------------------------------
SPX1_EGARCH11_mu_hat = zoo(matrix(, nrow=SPX1_roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(SPX1_EGARCH11_mu_hat) = c('SPX1_EGARCH11_mu_hat')
SPX1_EGARCH11_sigma_hat = zoo(matrix(, nrow=SPX1_roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(EGARCH11_sigma_hat) = c('EGARCH11_sigma_hat')
for (i in c(1:SPX1_roll))
{SPX1_EGARCH11 = GARCH_model_spec(mod="eGARCH", exreg=NULL)
try(withTimeout({(SPX1_EGARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPX1_EGARCH11))}, timeout = 5),silent = TRUE)
# the 247th value of R is 2004-12-31.
try((SPX1_EGARCH11fitforecast = GARCH_model_forecast(mod=SPX1_EGARCH11fit)), silent = TRUE)
# suppressWarnings(expr)
try((SPX1_EGARCH11_mu_hat[i]=SPX1_EGARCH11fitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPX1_EGARCH11_sigma_hat[i]=SPX1_EGARCH11fitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPX1_EGARCH11)
rm(SPX1_EGARCH11fit)
rm(SPX1_EGARCH11fitforecast)
}
# fitted(EGARCH11fitforecast)
SPX_RV_no_SPX1_EGARCH11_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPX1_roll)], as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))
-(na.omit(zoo((SPX1_EGARCH11_sigma_hat*0),
as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))))))
# Forecast Error, Mean and S.D.:
SPX1_EGARCH11forecasterror = (na.omit(SPX1_EGARCH11_sigma_hat)
-SPX_RV_no_SPX1_EGARCH11_sigma_hat_na_dated)
colnames(SPX1_EGARCH11forecasterror) = c("SPX1_EGARCH11forecasterror")
# plot(zoo(EGARCH11forecasterror, as.Date(row.names(EGARCH11forecasterror))), type='l', ylab='EGARCH11forecasterror', xlab='Date')
mean(SPX1_EGARCH11forecasterror)
sd(SPX1_EGARCH11forecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPX1_EGARCH11_sigma_hat_na_dated, (na.omit(SPX1_EGARCH11_sigma_hat)))
###----------------------------------------------------------------------------
### SPX1 In-sample SPX1_AR(1)-EGARCH(1,1)-SVI Model
###---------------------------------------------------------------------------
SPX1_in_sample_EGARCH11SVI = GARCH_model_spec(mod="eGARCH", exreg= as.matrix(SPX_dSVI_zoo[1:(246+1)]))
SPX1_in_sample_EGARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPX1_in_sample_EGARCH11SVI)
###----------------------------------------------------------------------------
### Create the SPX1_AR(1)-EGARCH(1,1)-SVI Model and forecasts
###----------------------------------------------------------------------------
SPX1_EGARCH11SVI_mu_hat = zoo(matrix(, nrow=SPX1_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(SPX1_EGARCH11SVI_mu_hat) = c('SPX1_EGARCH11SVI_mu_hat')
SPX1_EGARCH11SVI_sigma_hat = zoo(matrix(, nrow=SPX1_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPX1_roll)]))
colnames(EGARCH11SVI_sigma_hat) = c('EGARCH11SVI_sigma_hat')
for (i in c(1:SPX1_roll))
{SPX1_EGARCH11SVI = GARCH_model_spec(mod="eGARCH", exreg=as.matrix(SPX_dSVI_zoo[1:(246+i)]))
try(withTimeout({(SPX1_EGARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPX1_EGARCH11SVI))}, timeout = 5),silent = TRUE)
try((SPX1_EGARCH11SVIfitforecast = GARCH_model_forecast(mod=SPX1_EGARCH11SVIfit)), silent = TRUE)
try((SPX1_EGARCH11SVI_mu_hat[i]=SPX1_EGARCH11SVIfitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPX1_EGARCH11SVI_sigma_hat[i]=SPX1_EGARCH11SVIfitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPX1_EGARCH11SVI)
rm(SPX1_EGARCH11SVIfit)
rm(SPX1_EGARCH11SVIfitforecast)
}
SPX_RV_no_SPX1_EGARCH11SVI_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPX1_roll)], as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))
-(na.omit(zoo((SPX1_EGARCH11SVI_sigma_hat*0), as.Date(SPX_Dates[(248+1):(248+SPX1_roll)]))))))
# Forecast Error, Mean and S.D.:
SPX1_EGARCH11SVIforecasterror = (na.omit(SPX1_EGARCH11SVI_sigma_hat)
-SPX_RV_no_SPX1_EGARCH11SVI_sigma_hat_na_dated)
colnames(SPX1_EGARCH11SVIforecasterror) = c("SPX1_EGARCH11SVIforecasterror")
# plot(zoo(EGARCH11SVIforecasterror, as.Date(row.names(EGARCH11SVIforecasterror))), type='l', ylab='EGARCH11SVIforecasterror', xlab='Date')
mean(SPX1_EGARCH11SVIforecasterror)
sd(SPX1_EGARCH11SVIforecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPX1_EGARCH11SVI_sigma_hat_na_dated, (na.omit(SPX1_EGARCH11SVI_sigma_hat)))
###---------------------------------------------------------------------------
### s.d. forecase D-M tests
###---------------------------------------------------------------------------
print("This function implements the modified test proposed by Harvey, Leybourne and Newbold (1997). The null hypothesis is that the two methods have the same forecast accuracy. For alternative=less, the alternative hypothesis is that method 2 is less accurate than method 1. For alternative=greater, the alternative hypothesis is that method 2 is more accurate than method 1. For alternative=two.sided, the alternative hypothesis is that method 1 and method 2 have different levels of accuracy.")
print("Diebold and Mariano test SPX1_GARCH11 with and without SVI")
SPX1_GARCH11forecasterror_dm = SPX1_GARCH11forecasterror - (SPX1_GARCH11SVIforecasterror*0)
SPX1_GARCH11SVIforecasterror_dm = SPX1_GARCH11SVIforecasterror - (SPX1_GARCH11forecasterror*0)
SPX1_DM_Test_GARCH = dm.test(matrix(SPX1_GARCH11forecasterror_dm), matrix(SPX1_GARCH11SVIforecasterror_dm), alternative = "less")
print("Diebold and Mariano test SPX1_GJRGARCH11 with and without SVI")
SPX1_GJRGARCH11forecasterror_dm = SPX1_GJRGARCH11forecasterror - (SPX1_GJRGARCH11SVIforecasterror*0)
SPX1_GJRGARCH11SVIforecasterror_dm = SPX1_GJRGARCH11SVIforecasterror - (SPX1_GJRGARCH11forecasterror*0)
SPX1_DM_Test_GJRARCH = dm.test(matrix(SPX1_GJRGARCH11forecasterror_dm), matrix(SPX1_GJRGARCH11SVIforecasterror_dm), alternative = "less")
print("Diebold and Mariano test SPX1_EGARCH11 with and without SVI")
SPX1_EGARCH11forecasterror_dm = SPX1_EGARCH11forecasterror - (SPX1_EGARCH11SVIforecasterror*0)
SPX1_EGARCH11SVIforecasterror_dm = SPX1_EGARCH11SVIforecasterror - (SPX1_EGARCH11forecasterror*0)
SPX1_DM_Test_EGARCH = dm.test(matrix(SPX1_EGARCH11forecasterror_dm), matrix(SPX1_EGARCH11SVIforecasterror_dm), alternative = "less")
####---------------------------------------------------------------------------
#### SPX1 estimates of the probability of a positive return
####---------------------------------------------------------------------------
###----------------------------------------------------------------------------
### SPX1 C&D, Naive
###----------------------------------------------------------------------------
##-----------------------------------------------------------------------------
## Naive-Model and its forecast error statistics
##-----------------------------------------------------------------------------
# Naive-Model's Indicator function according to the GARCH11 Model:
SPX_Naive_I = ifelse(SPX_R_matrix>0 , 1 , 0)
# Naive Model:
SPX_Naive=(1:(length(SPX_R_matrix)))
for(t in c(1:(length(SPX_R_matrix))))
{SPX_Naive[t] = (1/t) * sum(SPX_R_matrix[1:t])}
# Note that the naive model provides the probability of aa positive return in
# the next time period and therefore spams from 2004-01-06 to 2019-03-13.
Naive_zoo = zoo(SPX_Naive, as.Date(rownames(as.matrix(SPX_R_zoo))))
##-----------------------------------------------------------------------------
## SPX1 C&D's GARCH models with and without SVI
## (Model independent variables' construction)
##-----------------------------------------------------------------------------
# mean of return up to time 'k'
SPX_R_mu=(1:length(SPX_R_matrix))
for (i in c(1:length(SPX_R_matrix))){SPX_R_mu[i]=mean(SPX_R_matrix[1:i])}
# standard deviation of return up to time 'k'.
# Note that its 1st value is NA since there is no standard deviation for a constant (according to R).
SPX_R_sigma=(1:length(SPX_R_matrix))
for (i in c(1:length(SPX_R_matrix))){R_sigma[i]=sd(SPX_R_matrix[1:i])}
SPX_R_sigma[1]=0 # Since the variance of a constant is 0.
#------------------------------------------------------------------------------
# SPX1 GARCH11's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GARCH11 Model:
SPX1_GARCH11_CandD_I = matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_GARCH11_CandD_I_k= matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_GARCH11_CandD_I_t = matrix( - (SPX1_GARCH11_mu_hat/SPX1_GARCH11_sigma_hat))
for(t in c(1:SPX1_roll))
{for (k in c(1:t))
{SPX1_GARCH11_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPX1_GARCH11_mu_hat[k])/SPX1_GARCH11_sigma_hat[k])
<=SPX1_GARCH11_CandD_I_t[t+1],
1,0)
# This will also be the number of out-of-sample predictions. N.B.: the 248th value of SPX_R_zoo corresponds to 2005-01-03, the 1st trading day out-of-sample, the first value after 252.
# Note that we have some missing values due to the model not always managing
# to converge to estimates of sigma and mu; since we need the t+1 model
# estimates to compute its CandD_I_k, we also loose its t'th value for
# each model's missing value.
# We also los the last value (the SPX1_roll'th value) for the same reason.
SPX1_GARCH11_CandD_I[k,t] =
ifelse((is.na(SPX1_GARCH11_CandD_I_k[k,t])), NA,
(sum(SPX1_GARCH11_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GARCH11 Model:
SPX1_GARCH11_CandD=(1:(SPX1_roll-1))
for(i in c(1:(SPX1_roll-1)))
{SPX1_GARCH11_CandD[i] = 1 - ((SPX1_GARCH11_CandD_I[i,i])/
length(na.omit(SPX1_GARCH11_CandD_I[1:i,i])))}
SPX1_GARCH11_CandD_zoo = zoo(SPX1_GARCH11_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPX1_roll-1))]))
# This zoo object has missing values.
SPX1_GARCH11_CandD_zoo_no_nas=zoo(na.omit(SPX1_GARCH11_CandD_zoo))
# rm(SPX1_GARCH11_CandD_I_t) # Cleaning datasets
# rm(SPX1_GARCH11_CandD_I_k)
# rm(SPX1_GARCH11_CandD_I)
#------------------------------------------------------------------------------
# SPX1 GARCH11SVI's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GARCH11SVI Model:
SPX1_GARCH11SVI_CandD_I = matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_GARCH11SVI_CandD_I_k= matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_GARCH11SVI_CandD_I_t = matrix( - (SPX1_GARCH11SVI_mu_hat/SPX1_GARCH11SVI_sigma_hat))
for(t in c(1:SPX1_roll))
{for (k in c(1:t))
{SPX1_GARCH11SVI_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPX1_GARCH11SVI_mu_hat[k])/SPX1_GARCH11SVI_sigma_hat[k])
<=SPX1_GARCH11SVI_CandD_I_t[t+1],
1,0)
SPX1_GARCH11SVI_CandD_I[k,t] =
ifelse((is.na(SPX1_GARCH11SVI_CandD_I_k[k,t])), NA,
(sum(SPX1_GARCH11SVI_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GARCH11SVI Model:
SPX1_GARCH11SVI_CandD=(1:(SPX1_roll-1))
for(i in c(1:(SPX1_roll-1)))
{SPX1_GARCH11SVI_CandD[i] = 1 - ((SPX1_GARCH11SVI_CandD_I[i,i])/
length(na.omit(SPX1_GARCH11SVI_CandD_I[1:i,i])))}
SPX1_GARCH11SVI_CandD_zoo = zoo(SPX1_GARCH11SVI_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPX1_roll-1))]))
# This zoo object has missing values.
SPX1_GARCH11SVI_CandD_zoo_no_nas=zoo(na.omit(SPX1_GARCH11SVI_CandD_zoo))
# rm(SPX1_GARCH11SVI_CandD_I_t) # Cleaning datasets
# rm(SPX1_GARCH11SVI_CandD_I_k)
# rm(SPX1_GARCH11SVI_CandD_I)
#------------------------------------------------------------------------------
# SPX1 GJRGARCH11's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GJRGARCH11 Model:
SPX1_GJRGARCH11_CandD_I = matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_GJRGARCH11_CandD_I_k= matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_GJRGARCH11_CandD_I_t = matrix( - (SPX1_GJRGARCH11_mu_hat/SPX1_GJRGARCH11_sigma_hat))
for(t in c(1:SPX1_roll))
{for (k in c(1:t))
{SPX1_GJRGARCH11_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPX1_GJRGARCH11_mu_hat[k])/SPX1_GJRGARCH11_sigma_hat[k])
<=SPX1_GJRGARCH11_CandD_I_t[t+1],
1,0)
# This will also be the number of out-of-sample predictions. N.B.: the 248th value of SPX_R_zoo corresponds to 2005-01-03, the 1st trading day out-of-sample, the first value after 252.
# Note that we have some missing values due to the model not always managing
# to converge to estimates of sigma and mu; since we need the t+1 model
# estimates to compute its CandD_I_k, we also loose its t'th value for
# each model's missing value.
# We also los the last value (the SPX1_roll'th value) for the same reason.
SPX1_GJRGARCH11_CandD_I[k,t] =
ifelse((is.na(SPX1_GJRGARCH11_CandD_I_k[k,t])), NA,
(sum(SPX1_GJRGARCH11_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GJRGARCH11 Model:
SPX1_GJRGARCH11_CandD=(1:(SPX1_roll-1))
for(i in c(1:(SPX1_roll-1)))
{SPX1_GJRGARCH11_CandD[i] = 1 - ((SPX1_GJRGARCH11_CandD_I[i,i])/
length(na.omit(SPX1_GJRGARCH11_CandD_I[1:i,i])))}
SPX1_GJRGARCH11_CandD_zoo = zoo(SPX1_GJRGARCH11_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPX1_roll-1))]))
# This zoo object has missing values.
SPX1_GJRGARCH11_CandD_zoo_no_nas=zoo(na.omit(SPX1_GJRGARCH11_CandD_zoo))
# rm(SPX1_GJRGARCH11_CandD_I_t) # Cleaning datasets
# rm(SPX1_GJRGARCH11_CandD_I_k)
# rm(SPX1_GJRGARCH11_CandD_I)
#------------------------------------------------------------------------------
# SPX1 GJRGARCH11SVI's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GJRGARCH11SVI Model:
SPX1_GJRGARCH11SVI_CandD_I = matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_GJRGARCH11SVI_CandD_I_k= matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_GJRGARCH11SVI_CandD_I_t = matrix( - (SPX1_GJRGARCH11SVI_mu_hat/SPX1_GJRGARCH11SVI_sigma_hat))
for(t in c(1:SPX1_roll))
{for (k in c(1:t))
{SPX1_GJRGARCH11SVI_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPX1_GJRGARCH11SVI_mu_hat[k])/SPX1_GJRGARCH11SVI_sigma_hat[k])
<=SPX1_GJRGARCH11SVI_CandD_I_t[t+1],
1,0)
SPX1_GJRGARCH11SVI_CandD_I[k,t] =
ifelse((is.na(SPX1_GJRGARCH11SVI_CandD_I_k[k,t])), NA,
(sum(SPX1_GJRGARCH11SVI_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GJRGARCH11SVI Model:
SPX1_GJRGARCH11SVI_CandD=(1:(SPX1_roll-1))
for(i in c(1:(SPX1_roll-1)))
{SPX1_GJRGARCH11SVI_CandD[i] = 1 - ((SPX1_GJRGARCH11SVI_CandD_I[i,i])/
length(na.omit(SPX1_GJRGARCH11SVI_CandD_I[1:i,i])))}
SPX1_GJRGARCH11SVI_CandD_zoo = zoo(SPX1_GJRGARCH11SVI_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPX1_roll-1))]))
# This zoo object has missing values.
SPX1_GJRGARCH11SVI_CandD_zoo_no_nas=zoo(na.omit(SPX1_GJRGARCH11SVI_CandD_zoo))
# rm(SPX1_GJRGARCH11SVI_CandD_I_t) # Cleaning datasets
# rm(SPX1_GJRGARCH11SVI_CandD_I_k)
# rm(SPX1_GJRGARCH11SVI_CandD_I)
#------------------------------------------------------------------------------
# SPX1 EGARCH11's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the EGARCH11 Model:
SPX1_EGARCH11_CandD_I = matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_EGARCH11_CandD_I_k= matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_EGARCH11_CandD_I_t = matrix( - (SPX1_EGARCH11_mu_hat/SPX1_EGARCH11_sigma_hat))
for(t in c(1:SPX1_roll))
{for (k in c(1:t))
{SPX1_EGARCH11_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPX1_EGARCH11_mu_hat[k])/SPX1_EGARCH11_sigma_hat[k])
<=SPX1_EGARCH11_CandD_I_t[t+1],
1,0)
# This will also be the number of out-of-sample predictions. N.B.: the 248th value of SPX_R_zoo corresponds to 2005-01-03, the 1st trading day out-of-sample, the first value after 252.
# Note that we have some missing values due to the model not always managing
# to converge to estimates of sigma and mu; since we need the t+1 model
# estimates to compute its CandD_I_k, we also loose its t'th value for
# each model's missing value.
# We also los the last value (the SPX1_roll'th value) for the same reason.
SPX1_EGARCH11_CandD_I[k,t] =
ifelse((is.na(SPX1_EGARCH11_CandD_I_k[k,t])), NA,
(sum(SPX1_EGARCH11_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the EGARCH11 Model:
SPX1_EGARCH11_CandD=(1:(SPX1_roll-1))
for(i in c(1:(SPX1_roll-1)))
{SPX1_EGARCH11_CandD[i] = 1 - ((SPX1_EGARCH11_CandD_I[i,i])/
length(na.omit(SPX1_EGARCH11_CandD_I[1:i,i])))}
SPX1_EGARCH11_CandD_zoo = zoo(SPX1_EGARCH11_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPX1_roll-1))]))
# This zoo object has missing values.
SPX1_EGARCH11_CandD_zoo_no_nas=zoo(na.omit(SPX1_EGARCH11_CandD_zoo))
# rm(SPX1_EGARCH11_CandD_I_t) # Cleaning datasets
# rm(SPX1_EGARCH11_CandD_I_k)
# rm(SPX1_EGARCH11_CandD_I)
#------------------------------------------------------------------------------
# SPX1 EGARCH11SVI's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the EGARCH11SVI Model:
SPX1_EGARCH11SVI_CandD_I = matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_EGARCH11SVI_CandD_I_k= matrix(, nrow=SPX1_roll, ncol=SPX1_roll)
SPX1_EGARCH11SVI_CandD_I_t = matrix( - (SPX1_EGARCH11SVI_mu_hat/SPX1_EGARCH11SVI_sigma_hat))
for(t in c(1:SPX1_roll))
{for (k in c(1:t))
{SPX1_EGARCH11SVI_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPX1_EGARCH11SVI_mu_hat[k])/SPX1_EGARCH11SVI_sigma_hat[k])
<=SPX1_EGARCH11SVI_CandD_I_t[t+1],
1,0)
SPX1_EGARCH11SVI_CandD_I[k,t] =
ifelse((is.na(SPX1_EGARCH11SVI_CandD_I_k[k,t])), NA,
(sum(SPX1_EGARCH11SVI_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the EGARCH11SVI Model:
SPX1_EGARCH11SVI_CandD=(1:(SPX1_roll-1))
for(i in c(1:(SPX1_roll-1)))
{SPX1_EGARCH11SVI_CandD[i] = 1 - ((SPX1_EGARCH11SVI_CandD_I[i,i])/
length(na.omit(SPX1_EGARCH11SVI_CandD_I[1:i,i])))}
SPX1_EGARCH11SVI_CandD_zoo = zoo(SPX1_EGARCH11SVI_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPX1_roll-1))]))
# This zoo object has missing values.
SPX1_EGARCH11SVI_CandD_zoo_no_nas=zoo(na.omit(SPX1_EGARCH11SVI_CandD_zoo))
# rm(SPX1_EGARCH11SVI_CandD_I_t) # Cleaning datasets
# rm(SPX1_EGARCH11SVI_CandD_I_k)
# rm(SPX1_EGARCH11SVI_CandD_I)
##-----------------------------------------------------------------------------
## SPX1 compare the predictive performance of each model
## (forecast error mean, sd, Brier Scores and Diebold&Mariano statistics)
##-----------------------------------------------------------------------------
# # mean of the probabilistic forecast errors derived from each model
SPX1_Observed_pi= ifelse(SPX_R_matrix>0,1,0)
SPX1_Observed_pi_zoo = zoo(SPX1_Observed_pi, as.Date(SPX_Dates))
SPX1_Naive_pi_error = SPX1_Naive_zoo - SPX1_Observed_pi_zoo
SPX1_GARCH11_pi_error = SPX1_GARCH11_CandD_zoo_no_nas - SPX1_Observed_pi_zoo
SPX1_GARCH11SVI_pi_error = SPX1_GARCH11SVI_CandD_zoo_no_nas - SPX1_Observed_pi_zoo
SPX1_GJRGARCH11_pi_error = SPX1_GJRGARCH11_CandD_zoo_no_nas - SPX1_Observed_pi_zoo
SPX1_GJRGARCH11SVI_pi_error = SPX1_GJRGARCH11SVI_CandD_zoo_no_nas - SPX1_Observed_pi_zoo
SPX1_EGARCH11_pi_error = SPX1_EGARCH11_CandD_zoo_no_nas - SPX1_Observed_pi_zoo
SPX1_EGARCH11SVI_pi_error = SPX1_EGARCH11SVI_CandD_zoo_no_nas - SPX1_Observed_pi_zoo
mean(SPX1_Naive_pi_error)
sd(SPX1_Naive_pi_error)
mean(SPX1_GARCH11_pi_error)
sd(SPX1_GARCH11_pi_error)
mean(SPX1_GARCH11SVI_pi_error)
sd(SPX1_GARCH11SVI_pi_error)
mean(SPX1_GJRGARCH11_pi_error)
sd(SPX1_GJRGARCH11_pi_error)
mean(SPX1_GJRGARCH11SVI_pi_error)
sd(SPX1_GJRGARCH11SVI_pi_error)
mean(SPX1_EGARCH11_pi_error)
sd(SPX1_EGARCH11_pi_error)
mean(SPX1_EGARCH11SVI_pi_error)
sd(SPX1_EGARCH11SVI_pi_error)
# # SPX1_Brier scores of the probabilistic forecast errors derived from each model
SPX1_Naive_pi_error_Brier_score =
(1/length(SPX1_Naive_pi_error))*sum(SPX1_Naive_pi_error^2)
show(SPX1_Naive_pi_error_Brier_score)
SPX1_GARCH11_pi_error_Brier_score =
(1/length(SPX1_GARCH11_pi_error))*sum(SPX1_GARCH11_pi_error^2)
show(SPX1_GARCH11_pi_error_Brier_score)
SPX1_GARCH11SVI_pi_error_Brier_score =
(1/length(SPX1_GARCH11SVI_pi_error))*sum(SPX1_GARCH11SVI_pi_error^2)
show(SPX1_GARCH11SVI_pi_error_Brier_score)
SPX1_GJRGARCH11_pi_error_Brier_score =
(1/length(SPX1_GJRGARCH11_pi_error))*sum(SPX1_GJRGARCH11_pi_error^2)
show(SPX1_GJRGARCH11_pi_error_Brier_score)
SPX1_GJRGARCH11SVI_pi_error_Brier_score =
(1/length(SPX1_GJRGARCH11SVI_pi_error))*sum(SPX1_GJRGARCH11SVI_pi_error^2)
show(SPX1_GJRGARCH11SVI_pi_error_Brier_score)
SPX1_EGARCH11_pi_error_Brier_score =
(1/length(SPX1_EGARCH11_pi_error))*sum(SPX1_EGARCH11_pi_error^2)
show(SPX1_EGARCH11_pi_error_Brier_score)
SPX1_EGARCH11SVI_pi_error_Brier_score =
(1/length(SPX1_EGARCH11SVI_pi_error))*sum(SPX1_EGARCH11SVI_pi_error^2)
show(SPX1_EGARCH11SVI_pi_error_Brier_score)
# # SPX1_Diebold&Mariano statistics
# Here the alternative hypothesis is that method 2 is more accurate than method 1; remember that a small p-value indicates strong evidence against the Null Hypothesis.
SPX1_GARCH11_pi_error_dm = SPX1_GARCH11_pi_error - (SPX1_GARCH11SVI_pi_error*0)
SPX1_GARCH11SVI_pi_error_dm = SPX1_GARCH11SVI_pi_error - (SPX1_GARCH11_pi_error*0)
SPX1_GARCH11_dm_test = dm.test(matrix(SPX1_GARCH11_pi_error_dm), matrix(SPX1_GARCH11SVI_pi_error_dm), alternative = "greater")
SPX1_GJRGARCH11_pi_error_dm = SPX1_GJRGARCH11_pi_error - (SPX1_GJRGARCH11SVI_pi_error*0)
SPX1_GJRGARCH11SVI_pi_error_dm = SPX1_GJRGARCH11SVI_pi_error - (SPX1_GJRGARCH11_pi_error*0)
SPX1_GJRGARCH11_dm_test = dm.test(matrix(SPX1_GJRGARCH11_pi_error_dm), matrix(SPX1_GJRGARCH11SVI_pi_error_dm), alternative = c("greater"))
SPX1_EGARCH11_pi_error_dm = SPX1_EGARCH11_pi_error - (SPX1_EGARCH11SVI_pi_error*0)
SPX1_EGARCH11SVI_pi_error_dm = SPX1_EGARCH11SVI_pi_error - (SPX1_EGARCH11_pi_error*0)
SPX1_EGARCH11_dm_test = dm.test(matrix(SPX1_EGARCH11_pi_error_dm), matrix(SPX1_EGARCH11SVI_pi_error_dm), alternative = c("greater"))
####----------------------------------------------------------------------------
#### SPX1's Financial significance
####----------------------------------------------------------------------------
# indicator of the realised direction of the return on the S&P 500 index
y_d = ifelse(SPX_R_zoo>0,1,0)
# Set the probability threashold at which to invest in the index in our 2d graphs:
for (p in c(0.49, 0.495, 0.5, 0.505, 0.51)){
###----------------------------------------------------------------------------
### SPX1 Granger and Pesaran (2000)'s framework using the GARCH11 model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPX1_GARCH11 = ifelse(SPX1_GARCH11_CandD_zoo_no_nas>p,1,0)
y_d_SPX1_GARCH11 = y_d - (SPX1_GARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GARCH11's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPX1_GARCH11),1))==1 , 1, 0)
R_Active_SPX1_GARCH11_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPX1_GARCH11*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPX1_GARCH11*0))))[2:length(y_hat_d_SPX1_GARCH11)]
R_Active_SPX1_GARCH11_p_cumulated = matrix(,nrow=length(R_Active_SPX1_GARCH11_p))
for (i in c(1:length(R_Active_SPX1_GARCH11_p)))
{R_Active_SPX1_GARCH11_p_cumulated[i] = prod((1+R_Active_SPX1_GARCH11_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPX1_GARCH11_p))
for (i in c(1:length(R_Active_SPX1_GARCH11_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPX1_GARCH11_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPX1 Granger and Pesaran (2000)'s framework using the GARCH11-SVI model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPX1_GARCH11SVI = ifelse(SPX1_GARCH11SVI_CandD_zoo_no_nas>p,1,0)
y_d_SPX1_GARCH11SVI = y_d - (SPX1_GARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GARCH11SVI's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPX1_GARCH11SVI),1))==1 , 1, 0)
R_Active_SPX1_GARCH11SVI_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPX1_GARCH11SVI*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPX1_GARCH11SVI*0))))[2:length(y_hat_d_SPX1_GARCH11SVI)]
R_Active_SPX1_GARCH11SVI_p_cumulated = matrix(,nrow=length(R_Active_SPX1_GARCH11SVI_p))
for (i in c(1:length(R_Active_SPX1_GARCH11SVI_p)))
{R_Active_SPX1_GARCH11SVI_p_cumulated[i] = prod((1+R_Active_SPX1_GARCH11SVI_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPX1_GARCH11SVI_p))
for (i in c(1:length(R_Active_SPX1_GARCH11SVI_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPX1_GARCH11SVI_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPX1 Granger and Pesaran (2000)'s framework using the GJRGARCH11 model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPX1_GJRGARCH11 = ifelse(SPX1_GJRGARCH11_CandD_zoo_no_nas>p,1,0)
y_d_SPX1_GJRGARCH11 = y_d - (SPX1_GJRGARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GJRGARCH11's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPX1_GJRGARCH11),1))==1 , 1, 0)
R_Active_SPX1_GJRGARCH11_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPX1_GJRGARCH11*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPX1_GJRGARCH11*0))))[2:length(y_hat_d_SPX1_GJRGARCH11)]
R_Active_SPX1_GJRGARCH11_p_cumulated = matrix(,nrow=length(R_Active_SPX1_GJRGARCH11_p))
for (i in c(1:length(R_Active_SPX1_GJRGARCH11_p)))
{R_Active_SPX1_GJRGARCH11_p_cumulated[i] = prod((1+R_Active_SPX1_GJRGARCH11_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPX1_GJRGARCH11_p))
for (i in c(1:length(R_Active_SPX1_GJRGARCH11_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPX1_GJRGARCH11_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPX1 Granger and Pesaran (2000)'s framework using the GJRGARCH11-SVI model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPX1_GJRGARCH11SVI = ifelse(SPX1_GJRGARCH11SVI_CandD_zoo_no_nas>p,1,0)
y_d_SPX1_GJRGARCH11SVI = y_d - (SPX1_GJRGARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GJRGARCH11SVI's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPX1_GJRGARCH11SVI),1))==1 , 1, 0)
R_Active_SPX1_GJRGARCH11SVI_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPX1_GJRGARCH11SVI*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPX1_GJRGARCH11SVI*0))))[2:length(y_hat_d_SPX1_GJRGARCH11SVI)]
R_Active_SPX1_GJRGARCH11SVI_p_cumulated = matrix(,nrow=length(R_Active_SPX1_GJRGARCH11SVI_p))
for (i in c(1:length(R_Active_SPX1_GJRGARCH11SVI_p)))
{R_Active_SPX1_GJRGARCH11SVI_p_cumulated[i] = prod((1+R_Active_SPX1_GJRGARCH11SVI_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPX1_GJRGARCH11SVI_p))
for (i in c(1:length(R_Active_SPX1_GJRGARCH11SVI_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPX1_GJRGARCH11SVI_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPX1 Granger and Pesaran (2000)'s framework using the EGARCH11 model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPX1_EGARCH11 = ifelse(SPX1_EGARCH11_CandD_zoo_no_nas>p,1,0)
y_d_SPX1_EGARCH11 = y_d - (SPX1_EGARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to EGARCH11's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPX1_EGARCH11),1))==1 , 1, 0)
R_Active_SPX1_EGARCH11_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPX1_EGARCH11*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPX1_EGARCH11*0))))[2:length(y_hat_d_SPX1_EGARCH11)]
R_Active_SPX1_EGARCH11_p_cumulated = matrix(,nrow=length(R_Active_SPX1_EGARCH11_p))
for (i in c(1:length(R_Active_SPX1_EGARCH11_p)))
{R_Active_SPX1_EGARCH11_p_cumulated[i] = prod((1+R_Active_SPX1_EGARCH11_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPX1_EGARCH11_p))
for (i in c(1:length(R_Active_SPX1_EGARCH11_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPX1_EGARCH11_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPX1 Granger and Pesaran (2000)'s framework using the EGARCH11-SVI model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPX1_EGARCH11SVI = ifelse(SPX1_EGARCH11SVI_CandD_zoo_no_nas>p,1,0)
y_d_SPX1_EGARCH11SVI = y_d - (SPX1_EGARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to EGARCH11SVI's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPX1_EGARCH11SVI),1))==1 , 1, 0)
R_Active_SPX1_EGARCH11SVI_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPX1_EGARCH11SVI*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPX1_EGARCH11SVI*0))))[2:length(y_hat_d_SPX1_EGARCH11SVI)]
R_Active_SPX1_EGARCH11SVI_p_cumulated = matrix(,nrow=length(R_Active_SPX1_EGARCH11SVI_p))
for (i in c(1:length(R_Active_SPX1_EGARCH11SVI_p)))
{R_Active_SPX1_EGARCH11SVI_p_cumulated[i] = prod((1+R_Active_SPX1_EGARCH11SVI_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPX1_EGARCH11SVI_p))
for (i in c(1:length(R_Active_SPX1_EGARCH11SVI_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPX1_EGARCH11SVI_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPX1 Granger and Pesaran (2000)'s framework 2d Graphs put together
###----------------------------------------------------------------------------
# Plot all 2d polts
plot(zoo(R_Active_SPX1_GARCH11_p_cumulated,
as.Date(zoo::index(y_hat_d_SPX1_GARCH11[2:length(y_hat_d_SPX1_GARCH11)]))),
cex.axis=1, type="l",col="red", xlab='Date', ylim=c(0,3), ylab='SPX1 strategy gain ($)', main=str_c("SPX1 Strategy for psi ", p))
lines(zoo(R_cumulated,
as.Date(zoo::index(y_hat_d_SPX1_GARCH11[2:length(y_hat_d_SPX1_GARCH11)]))),
col="black")
lines(zoo(R_Active_SPX1_GARCH11SVI_p_cumulated, as.Date(zoo::index(y_hat_d_SPX1_GARCH11SVI[2:length(y_hat_d_SPX1_GARCH11SVI)]))),
col="orange")
lines(zoo(R_Active_SPX1_GJRGARCH11_p_cumulated,
as.Date(zoo::index(y_hat_d_SPX1_GJRGARCH11[2:length(y_hat_d_SPX1_GJRGARCH11)]))),
col="blue")
lines(zoo(R_Active_SPX1_GJRGARCH11SVI_p_cumulated,
as.Date(zoo::index(y_hat_d_SPX1_GJRGARCH11SVI[2:length(y_hat_d_SPX1_GJRGARCH11SVI)]))),
col="magenta")
lines(zoo(R_Active_SPX1_EGARCH11_p_cumulated,
as.Date(zoo::index(y_hat_d_SPX1_EGARCH11[2:length(y_hat_d_SPX1_EGARCH11)]))),
col="green")
lines(zoo(R_Active_SPX1_EGARCH11SVI_p_cumulated,
as.Date(zoo::index(y_hat_d_SPX1_EGARCH11SVI[2:length(y_hat_d_SPX1_EGARCH11SVI)]))),
col="purple")
legend(lty=1, cex=1,
"topleft", col=c("black", "red", "orange", "blue", "magenta", "green", "purple"),
legend=c("Buy and hold", "GARCH", "GARCH-SVI", "GJRGARCH", "GJRGARCH-SVI", "EGARCH", "EGARCH-SVI"))
}
####---------------------------------------------------------------------------
#### SPXCPV GARCH models: create, train/fit them and create forecasts
####---------------------------------------------------------------------------
# Set parameters
SPXCPV_roll = length(SPX_dSVICPV)-247
# This will also be the number of out-of-sample predictions. N.B.: the 248th
# value of R corresponds to 2005-01-03, the 1st trading day out-of-sample,
# the first value after 247.
###----------------------------------------------------------------------------
### SPXCPV In-sample SPXCPV_AR(1)-GARCH(1,1) Model
###---------------------------------------------------------------------------
SPXCPV_in_sample_GARCH11 = GARCH_model_spec(mod="sGARCH", exreg= NULL)
SPXCPV_in_sample_GARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPXCPV_in_sample_GARCH11)
###----------------------------------------------------------------------------
### Create the SPXCPV_AR(1)-GARCH(1,1) Model and forecasts
###----------------------------------------------------------------------------
SPXCPV_GARCH11_mu_hat = zoo(matrix(, nrow=roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(SPXCPV_GARCH11_mu_hat) = c('SPXCPV_GARCH11_mu_hat')
SPXCPV_GARCH11_sigma_hat = zoo(matrix(, nrow=roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(SPXCPV_GARCH11_sigma_hat) = c('SPXCPV_GARCH11_sigma_hat')
for (i in c(1:SPXCPV_roll))
{SPXCPV_GARCH11 = GARCH_model_spec(mod="sGARCH", exreg=NULL)
try(withTimeout({(SPXCPV_GARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPXCPV_GARCH11))}, timeout = 5),silent = TRUE)
# the 247th value of R is 2004-12-31.
try((SPXCPV_GARCH11fitforecast = GARCH_model_forecast(mod=SPXCPV_GARCH11fit)), silent = TRUE)
# suppressWarnings(expr)
try((SPXCPV_GARCH11_mu_hat[i]=SPXCPV_GARCH11fitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPXCPV_GARCH11_sigma_hat[i]=SPXCPV_GARCH11fitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPXCPV_GARCH11)
rm(SPXCPV_GARCH11fit)
rm(SPXCPV_GARCH11fitforecast)
}
SPX_RV_no_SPXCPV_GARCH11_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPXCPV_roll)], as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))
-(na.omit(zoo((SPXCPV_GARCH11_sigma_hat*0),
as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))))))
# Forecast Error, Mean and S.D.:
SPXCPV_GARCH11forecasterror = (na.omit(SPXCPV_GARCH11_sigma_hat)
-SPX_RV_no_SPXCPV_GARCH11_sigma_hat_na_dated)
colnames(SPXCPV_GARCH11forecasterror) = c("SPXCPV_GARCH11forecasterror")
mean(SPXCPV_GARCH11forecasterror)
sd(SPXCPV_GARCH11forecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPXCPV_GARCH11_sigma_hat_na_dated, (na.omit(SPXCPV_GARCH11_sigma_hat)))
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.1_SPX_CPV_GARCH11.RData")
###----------------------------------------------------------------------------
### SPXCPV In-sample SPXCPV_AR(1)-GARCH(1,1)-SVI Model
###---------------------------------------------------------------------------
SPXCPV_in_sample_GARCH11SVI = GARCH_model_spec(mod="sGARCH", exreg= as.matrix(SPX_dSVICPV[1:(246+1)]))
SPXCPV_in_sample_GARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPXCPV_in_sample_GARCH11SVI)
###----------------------------------------------------------------------------
### Create the SPXCPV_AR(1)-GARCH(1,1)-SVI Model and forecasts
###----------------------------------------------------------------------------
SPXCPV_GARCH11SVI_mu_hat = zoo(matrix(, nrow=SPXCPV_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(SPXCPV_GARCH11SVI_mu_hat) = c('SPXCPV_GARCH11_mu_hat')
SPXCPV_GARCH11SVI_sigma_hat = zoo(matrix(, nrow=SPXCPV_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(GARCH11_sigma_hat) = c('GARCH11_sigma_hat')
for (i in c(1:SPXCPV_roll))
{SPXCPV_GARCH11SVI = GARCH_model_spec(mod="sGARCH", exreg=as.matrix(SPX_dSVICPV[1:(246+i)]))
try(withTimeout({(SPXCPV_GARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPXCPV_GARCH11SVI))}, timeout = 5),silent = TRUE)
try((SPXCPV_GARCH11SVIfitforecast = GARCH_model_forecast(mod=SPXCPV_GARCH11SVIfit)), silent = TRUE)
try((SPXCPV_GARCH11SVI_mu_hat[i]=SPXCPV_GARCH11SVIfitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPXCPV_GARCH11SVI_sigma_hat[i]=SPXCPV_GARCH11SVIfitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPXCPV_GARCH11SVI)
rm(SPXCPV_GARCH11SVIfit)
rm(SPXCPV_GARCH11SVIfitforecast)
}
# fitted(GARCH11SVIfitforecast)
SPX_RV_no_SPXCPV_GARCH11SVI_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPXCPV_roll)], as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))
-(na.omit(zoo((SPXCPV_GARCH11SVI_sigma_hat*0), as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))))))
# Forecast Error, Mean and S.D.:
SPXCPV_GARCH11SVIforecasterror = (na.omit(SPXCPV_GARCH11SVI_sigma_hat)
-SPX_RV_no_SPXCPV_GARCH11SVI_sigma_hat_na_dated)
colnames(SPXCPV_GARCH11SVIforecasterror) = c("SPXCPV_GARCH11SVIforecasterror")
# plot(zoo(GARCH11SVIforecasterror, as.Date(row.names(GARCH11SVIforecasterror))), type='l', ylab='GARCH11SVIforecasterror', xlab='Date')
mean(SPXCPV_GARCH11SVIforecasterror)
sd(SPXCPV_GARCH11SVIforecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPXCPV_GARCH11SVI_sigma_hat_na_dated, (na.omit(SPXCPV_GARCH11SVI_sigma_hat)))
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.2_SPX_CPV_GARCH11SVI.RData")
###----------------------------------------------------------------------------
### SPXCPV In-sample SPXCPV_AR(1)-GJRGARCH(1,1) Model
###---------------------------------------------------------------------------
SPXCPV_in_sample_GJRGARCH11 = GARCH_model_spec(mod="gjrGARCH", exreg= NULL)
SPXCPV_in_sample_GJRGARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPXCPV_in_sample_GJRGARCH11)
###----------------------------------------------------------------------------
### Create the SPXCPV_AR(1)-GJRGARCH(1,1) Model and forecasts
###----------------------------------------------------------------------------
SPXCPV_GJRGARCH11_mu_hat = zoo(matrix(, nrow=SPXCPV_roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(SPXCPV_GJRGARCH11_mu_hat) = c('SPXCPV_GJRGARCH11_mu_hat')
SPXCPV_GJRGARCH11_sigma_hat = zoo(matrix(, nrow=SPXCPV_roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(GJRGARCH11_sigma_hat) = c('GJRGARCH11_sigma_hat')
for (i in c(1:SPXCPV_roll))
{SPXCPV_GJRGARCH11 = GARCH_model_spec(mod="gjrGARCH", exreg=NULL)
try(withTimeout({(SPXCPV_GJRGARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPXCPV_GJRGARCH11))}, timeout = 5),silent = TRUE)
# the 247th value of R is 2004-12-31.
try((SPXCPV_GJRGARCH11fitforecast = GARCH_model_forecast(mod=SPXCPV_GJRGARCH11fit)), silent = TRUE)
# suppressWarnings(expr)
try((SPXCPV_GJRGARCH11_mu_hat[i]=SPXCPV_GJRGARCH11fitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPXCPV_GJRGARCH11_sigma_hat[i]=SPXCPV_GJRGARCH11fitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPXCPV_GJRGARCH11)
rm(SPXCPV_GJRGARCH11fit)
rm(SPXCPV_GJRGARCH11fitforecast)
}
# fitted(GJRGARCH11fitforecast)
SPX_RV_no_SPXCPV_GJRGARCH11_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPXCPV_roll)], as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))
-(na.omit(zoo((SPXCPV_GJRGARCH11_sigma_hat*0),
as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))))))
# Forecast Error, Mean and S.D.:
SPXCPV_GJRGARCH11forecasterror = (na.omit(SPXCPV_GJRGARCH11_sigma_hat)
-SPX_RV_no_SPXCPV_GJRGARCH11_sigma_hat_na_dated)
colnames(SPXCPV_GJRGARCH11forecasterror) = c("SPXCPV_GJRGARCH11forecasterror")
# plot(zoo(GJRGARCH11forecasterror, as.Date(row.names(GJRGARCH11forecasterror))), type='l', ylab='GJRGARCH11forecasterror', xlab='Date')
mean(SPXCPV_GJRGARCH11forecasterror)
sd(SPXCPV_GJRGARCH11forecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPXCPV_GJRGARCH11_sigma_hat_na_dated, (na.omit(SPXCPV_GJRGARCH11_sigma_hat)))
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.3_SPX_CPV_GJRGARCH11.RData")
###----------------------------------------------------------------------------
### SPXCPV In-sample SPXCPV_AR(1)-GJRGARCH(1,1)-SVI Model
###---------------------------------------------------------------------------
SPXCPV_in_sample_GJRGARCH11SVI = GARCH_model_spec(mod="gjrGARCH", exreg= as.matrix(SPX_dSVICPV[1:(246+1)]))
SPXCPV_in_sample_GJRGARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPXCPV_in_sample_GJRGARCH11SVI)
###----------------------------------------------------------------------------
### Create the SPXCPV_AR(1)-GJRGARCH(1,1)-SVI Model and forecasts
###----------------------------------------------------------------------------
SPXCPV_GJRGARCH11SVI_mu_hat = zoo(matrix(, nrow=SPXCPV_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(SPXCPV_GJRGARCH11SVI_mu_hat) = c('SPXCPV_GJRGARCH11_mu_hat')
SPXCPV_GJRGARCH11SVI_sigma_hat = zoo(matrix(, nrow=SPXCPV_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(GJRGARCH11_sigma_hat) = c('GJRGARCH11_sigma_hat')
for (i in c(1:SPXCPV_roll))
{SPXCPV_GJRGARCH11SVI = GARCH_model_spec(mod="gjrGARCH", exreg=as.matrix(SPX_dSVICPV[1:(246+i)]))
try(withTimeout({(SPXCPV_GJRGARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPXCPV_GJRGARCH11SVI))}, timeout = 5),silent = TRUE)
try((SPXCPV_GJRGARCH11SVIfitforecast = GARCH_model_forecast(mod=SPXCPV_GJRGARCH11SVIfit)), silent = TRUE)
try((SPXCPV_GJRGARCH11SVI_mu_hat[i]=SPXCPV_GJRGARCH11SVIfitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPXCPV_GJRGARCH11SVI_sigma_hat[i]=SPXCPV_GJRGARCH11SVIfitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPXCPV_GJRGARCH11SVI)
rm(SPXCPV_GJRGARCH11SVIfit)
rm(SPXCPV_GJRGARCH11SVIfitforecast)
}
SPX_RV_no_SPXCPV_GJRGARCH11SVI_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPXCPV_roll)], as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))
-(na.omit(zoo((SPXCPV_GJRGARCH11SVI_sigma_hat*0), as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))))))
# Forecast Error, Mean and S.D.:
SPXCPV_GJRGARCH11SVIforecasterror = (na.omit(SPXCPV_GJRGARCH11SVI_sigma_hat)
-SPX_RV_no_SPXCPV_GJRGARCH11SVI_sigma_hat_na_dated)
colnames(SPXCPV_GJRGARCH11SVIforecasterror) = c("SPXCPV_GJRGARCH11SVIforecasterror")
# plot(zoo(GJRGARCH11SVIforecasterror, as.Date(row.names(GJRGARCH11SVIforecasterror))), type='l', ylab='GJRGARCH11SVIforecasterror', xlab='Date')
mean(SPXCPV_GJRGARCH11SVIforecasterror)
sd(SPXCPV_GJRGARCH11SVIforecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPXCPV_GJRGARCH11SVI_sigma_hat_na_dated, (na.omit(SPXCPV_GJRGARCH11SVI_sigma_hat)))
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.4_SPX_CPV_GJRGARCH11SVI.RData")
###----------------------------------------------------------------------------
### SPXCPV In-sample SPXCPV_AR(1)-EGARCH(1,1) Model
###---------------------------------------------------------------------------
SPXCPV_in_sample_EGARCH11 = GARCH_model_spec(mod="eGARCH", exreg= NULL)
SPXCPV_in_sample_EGARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPXCPV_in_sample_EGARCH11)
###----------------------------------------------------------------------------
### Create the SPXCPV_AR(1)-EGARCH(1,1) Model and forecasts
###----------------------------------------------------------------------------
SPXCPV_EGARCH11_mu_hat = zoo(matrix(, nrow=SPXCPV_roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(SPXCPV_EGARCH11_mu_hat) = c('SPXCPV_EGARCH11_mu_hat')
SPXCPV_EGARCH11_sigma_hat = zoo(matrix(, nrow=SPXCPV_roll, ncol=1),
as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(EGARCH11_sigma_hat) = c('EGARCH11_sigma_hat')
for (i in c(1:SPXCPV_roll))
{SPXCPV_EGARCH11 = GARCH_model_spec(mod="eGARCH", exreg=NULL)
try(withTimeout({(SPXCPV_EGARCH11fit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPXCPV_EGARCH11))}, timeout = 5),silent = TRUE)
# the 247th value of R is 2004-12-31.
try((SPXCPV_EGARCH11fitforecast = GARCH_model_forecast(mod=SPXCPV_EGARCH11fit)), silent = TRUE)
# suppressWarnings(expr)
try((SPXCPV_EGARCH11_mu_hat[i]=SPXCPV_EGARCH11fitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPXCPV_EGARCH11_sigma_hat[i]=SPXCPV_EGARCH11fitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPXCPV_EGARCH11)
rm(SPXCPV_EGARCH11fit)
rm(SPXCPV_EGARCH11fitforecast)
}
# fitted(EGARCH11fitforecast)
SPX_RV_no_SPXCPV_EGARCH11_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPXCPV_roll)], as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))
-(na.omit(zoo((SPXCPV_EGARCH11_sigma_hat*0),
as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))))))
# Forecast Error, Mean and S.D.:
SPXCPV_EGARCH11forecasterror = (na.omit(SPXCPV_EGARCH11_sigma_hat)
-SPX_RV_no_SPXCPV_EGARCH11_sigma_hat_na_dated)
colnames(SPXCPV_EGARCH11forecasterror) = c("SPXCPV_EGARCH11forecasterror")
# plot(zoo(EGARCH11forecasterror, as.Date(row.names(EGARCH11forecasterror))), type='l', ylab='EGARCH11forecasterror', xlab='Date')
mean(SPXCPV_EGARCH11forecasterror)
sd(SPXCPV_EGARCH11forecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPXCPV_EGARCH11_sigma_hat_na_dated, (na.omit(SPXCPV_EGARCH11_sigma_hat)))
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.5_SPX_CPV_EGARCH11.RData")
###----------------------------------------------------------------------------
### SPXCPV In-sample SPXCPV_AR(1)-EGARCH(1,1)-SVI Model
###---------------------------------------------------------------------------
SPXCPV_in_sample_EGARCH11SVI = GARCH_model_spec(mod="eGARCH", exreg= as.matrix(SPX_dSVICPV[1:(246+1)]))
SPXCPV_in_sample_EGARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+1)],
spec=SPXCPV_in_sample_EGARCH11SVI)
###----------------------------------------------------------------------------
### Create the SPXCPV_AR(1)-EGARCH(1,1)-SVI Model and forecasts
###----------------------------------------------------------------------------
SPXCPV_EGARCH11SVI_mu_hat = zoo(matrix(, nrow=SPXCPV_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(SPXCPV_EGARCH11SVI_mu_hat) = c('SPXCPV_EGARCH11SVI_mu_hat')
SPXCPV_EGARCH11SVI_sigma_hat = zoo(matrix(, nrow=SPXCPV_roll, ncol=1), as.Date(SPX_Datesm1[(247+1):(247+SPXCPV_roll)]))
colnames(EGARCH11SVI_sigma_hat) = c('EGARCH11SVI_sigma_hat')
for (i in c(1:SPXCPV_roll))
{SPXCPV_EGARCH11SVI = GARCH_model_spec(mod="eGARCH", exreg=as.matrix(SPX_dSVICPV[1:(246+i)]))
try(withTimeout({(SPXCPV_EGARCH11SVIfit = ugarchfit(data = SPX_R_matrix[1:(246+i)], spec=SPXCPV_EGARCH11SVI))}, timeout = 5),silent = TRUE)
try((SPXCPV_EGARCH11SVIfitforecast = GARCH_model_forecast(mod=SPXCPV_EGARCH11SVIfit)), silent = TRUE)
try((SPXCPV_EGARCH11SVI_mu_hat[i]=SPXCPV_EGARCH11SVIfitforecast@forecast[["seriesFor"]]),silent = TRUE)
try((SPXCPV_EGARCH11SVI_sigma_hat[i]=SPXCPV_EGARCH11SVIfitforecast@forecast[["sigmaFor"]]), silent = TRUE)
rm(SPXCPV_EGARCH11SVI)
rm(SPXCPV_EGARCH11SVIfit)
rm(SPXCPV_EGARCH11SVIfitforecast)
}
SPXCPV_EGARCH11SVI_sigma_hat[2874] = NA # This is due to a bad conversion specifically for EGARCH11SVI at i=2874.
SPX_RV_no_SPXCPV_EGARCH11SVI_sigma_hat_na_dated = as.matrix(
(zoo(SPX_RV[(248+1):(248+SPXCPV_roll)], as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))
-(na.omit(zoo((SPXCPV_EGARCH11SVI_sigma_hat*0), as.Date(SPX_Dates[(248+1):(248+SPXCPV_roll)]))))))
# Forecast Error, Mean and S.D.:
SPXCPV_EGARCH11SVIforecasterror = (na.omit(SPXCPV_EGARCH11SVI_sigma_hat)
-SPX_RV_no_SPXCPV_EGARCH11SVI_sigma_hat_na_dated)
colnames(SPXCPV_EGARCH11SVIforecasterror) = c("SPXCPV_EGARCH11SVIforecasterror")
# plot(zoo(EGARCH11SVIforecasterror, as.Date(row.names(EGARCH11SVIforecasterror))), type='l', ylab='EGARCH11SVIforecasterror', xlab='Date')
mean(SPXCPV_EGARCH11SVIforecasterror)
sd(SPXCPV_EGARCH11SVIforecasterror)
# RMSE of the sigme (standard deviations) of the forecast:
rmse(SPX_RV_no_SPXCPV_EGARCH11SVI_sigma_hat_na_dated, (na.omit(SPXCPV_EGARCH11SVI_sigma_hat)))
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.6_SPX_CPV_EGARCH11SVI.RData")
###---------------------------------------------------------------------------
### s.d. forecase D-M tests
###---------------------------------------------------------------------------
print("This function implements the modified test proposed by Harvey, Leybourne and Newbold (1997). The null hypothesis is that the two methods have the same forecast accuracy. For alternative=less, the alternative hypothesis is that method 2 is less accurate than method 1. For alternative=greater, the alternative hypothesis is that method 2 is more accurate than method 1. For alternative=two.sided, the alternative hypothesis is that method 1 and method 2 have different levels of accuracy.")
print("Diebold and Mariano test SPXCPV_GARCH11 with and without SVI")
SPXCPV_GARCH11forecasterror_dm = SPXCPV_GARCH11forecasterror - (SPXCPV_GARCH11SVIforecasterror*0)
SPXCPV_GARCH11SVIforecasterror_dm = SPXCPV_GARCH11SVIforecasterror - (SPXCPV_GARCH11forecasterror*0)
SPXCPV_DM_Test_GARCH = dm.test(matrix(SPXCPV_GARCH11forecasterror_dm), matrix(SPXCPV_GARCH11SVIforecasterror_dm), alternative = "less")
print("Diebold and Mariano test SPXCPV_GJRGARCH11 with and without SVI")
SPXCPV_GJRGARCH11forecasterror_dm = SPXCPV_GJRGARCH11forecasterror - (SPXCPV_GJRGARCH11SVIforecasterror*0)
SPXCPV_GJRGARCH11SVIforecasterror_dm = SPXCPV_GJRGARCH11SVIforecasterror - (SPXCPV_GJRGARCH11forecasterror*0)
SPXCPV_DM_Test_GJRARCH = dm.test(matrix(SPXCPV_GJRGARCH11forecasterror_dm), matrix(SPXCPV_GJRGARCH11SVIforecasterror_dm), alternative = "less")
print("Diebold and Mariano test SPXCPV_EGARCH11 with and without SVI")
SPXCPV_EGARCH11forecasterror_dm = SPXCPV_EGARCH11forecasterror - (SPXCPV_EGARCH11SVIforecasterror*0)
SPXCPV_EGARCH11SVIforecasterror_dm = SPXCPV_EGARCH11SVIforecasterror - (SPXCPV_EGARCH11forecasterror*0)
SPXCPV_DM_Test_EGARCH = dm.test(matrix(SPXCPV_EGARCH11forecasterror_dm), matrix(SPXCPV_EGARCH11SVIforecasterror_dm), alternative = "less")
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.7_SPX_CPV_DM.RData")
####---------------------------------------------------------------------------
#### SPXCPV estimates of the probability of a positive return
####---------------------------------------------------------------------------
###----------------------------------------------------------------------------
### C&D, Naive
###----------------------------------------------------------------------------
#------------------------------------------------------------------------------
# SPXCPV GARCH11's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GARCH11 Model:
SPXCPV_GARCH11_CandD_I = matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_GARCH11_CandD_I_k= matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_GARCH11_CandD_I_t = matrix( - (SPXCPV_GARCH11_mu_hat/SPXCPV_GARCH11_sigma_hat))
for(t in c(1:SPXCPV_roll))
{for (k in c(1:t))
{SPXCPV_GARCH11_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPXCPV_GARCH11_mu_hat[k])/SPXCPV_GARCH11_sigma_hat[k])
<=SPXCPV_GARCH11_CandD_I_t[t+1],
1,0)
SPXCPV_GARCH11_CandD_I[k,t] =
ifelse((is.na(SPXCPV_GARCH11_CandD_I_k[k,t])), NA,
(sum(SPXCPV_GARCH11_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GARCH11 Model:
SPXCPV_GARCH11_CandD=(1:(SPXCPV_roll-1))
for(i in c(1:(SPXCPV_roll-1)))
{SPXCPV_GARCH11_CandD[i] = 1 - ((SPXCPV_GARCH11_CandD_I[i,i])/
length(na.omit(SPXCPV_GARCH11_CandD_I[1:i,i])))}
SPXCPV_GARCH11_CandD_zoo = zoo(SPXCPV_GARCH11_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPXCPV_roll-1))]))
# This zoo object has missing values.
SPXCPV_GARCH11_CandD_zoo_no_nas=zoo(na.omit(SPXCPV_GARCH11_CandD_zoo))
# rm(SPXCPV_GARCH11_CandD_I_t) # Cleaning datasets
# rm(SPXCPV_GARCH11_CandD_I_k)
# rm(SPXCPV_GARCH11_CandD_I)
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.8_SPX_CPV_GARCH11_CandD.RData")
#------------------------------------------------------------------------------
# SPXCPV GARCH11SVI's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GARCH11SVI Model:
SPXCPV_GARCH11SVI_CandD_I = matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_GARCH11SVI_CandD_I_k= matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_GARCH11SVI_CandD_I_t = matrix( - (SPXCPV_GARCH11SVI_mu_hat/SPXCPV_GARCH11SVI_sigma_hat))
for(t in c(1:SPXCPV_roll))
{for (k in c(1:t))
{SPXCPV_GARCH11SVI_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPXCPV_GARCH11SVI_mu_hat[k])/SPXCPV_GARCH11SVI_sigma_hat[k])
<=SPXCPV_GARCH11SVI_CandD_I_t[t+1],
1,0)
SPXCPV_GARCH11SVI_CandD_I[k,t] =
ifelse((is.na(SPXCPV_GARCH11SVI_CandD_I_k[k,t])), NA,
(sum(SPXCPV_GARCH11SVI_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GARCH11SVI Model:
SPXCPV_GARCH11SVI_CandD=(1:(SPXCPV_roll-1))
for(i in c(1:(SPXCPV_roll-1)))
{SPXCPV_GARCH11SVI_CandD[i] = 1 - ((SPXCPV_GARCH11SVI_CandD_I[i,i])/
length(na.omit(SPXCPV_GARCH11SVI_CandD_I[1:i,i])))}
SPXCPV_GARCH11SVI_CandD_zoo = zoo(SPXCPV_GARCH11SVI_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPXCPV_roll-1))]))
# This zoo object has missing values.
SPXCPV_GARCH11SVI_CandD_zoo_no_nas=zoo(na.omit(SPXCPV_GARCH11SVI_CandD_zoo))
# rm(SPXCPV_GARCH11SVI_CandD_I_t) # Cleaning datasets
# rm(SPXCPV_GARCH11SVI_CandD_I_k)
# rm(SPXCPV_GARCH11SVI_CandD_I)
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.9_SPX_CPV_GARCH11SVI_CandD.RData")
#------------------------------------------------------------------------------
# SPXCPV GJRGARCH11's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GJRGARCH11 Model:
SPXCPV_GJRGARCH11_CandD_I = matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_GJRGARCH11_CandD_I_k= matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_GJRGARCH11_CandD_I_t = matrix( - (SPXCPV_GJRGARCH11_mu_hat/SPXCPV_GJRGARCH11_sigma_hat))
for(t in c(1:SPXCPV_roll))
{for (k in c(1:t))
{SPXCPV_GJRGARCH11_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPXCPV_GJRGARCH11_mu_hat[k])/SPXCPV_GJRGARCH11_sigma_hat[k])
<=SPXCPV_GJRGARCH11_CandD_I_t[t+1],
1,0)
# This will also be the number of out-of-sample predictions. N.B.: the 248th value of SPX_R_zoo corresponds to 2005-01-03, the 1st trading day out-of-sample, the first value after 252.
# Note that we have some missing values due to the model not always managing
# to converge to estimates of sigma and mu; since we need the t+1 model
# estimates to compute its CandD_I_k, we also loose its t'th value for
# each model's missing value.
# We also los the last value (the SPXCPV_roll'th value) for the same reason.
SPXCPV_GJRGARCH11_CandD_I[k,t] =
ifelse((is.na(SPXCPV_GJRGARCH11_CandD_I_k[k,t])), NA,
(sum(SPXCPV_GJRGARCH11_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GJRGARCH11 Model:
SPXCPV_GJRGARCH11_CandD=(1:(SPXCPV_roll-1))
for(i in c(1:(SPXCPV_roll-1)))
{SPXCPV_GJRGARCH11_CandD[i] = 1 - ((SPXCPV_GJRGARCH11_CandD_I[i,i])/
length(na.omit(SPXCPV_GJRGARCH11_CandD_I[1:i,i])))}
SPXCPV_GJRGARCH11_CandD_zoo = zoo(SPXCPV_GJRGARCH11_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPXCPV_roll-1))]))
# This zoo object has missing values.
SPXCPV_GJRGARCH11_CandD_zoo_no_nas=zoo(na.omit(SPXCPV_GJRGARCH11_CandD_zoo))
# rm(SPXCPV_GJRGARCH11_CandD_I_t) # Cleaning datasets
# rm(SPXCPV_GJRGARCH11_CandD_I_k)
# rm(SPXCPV_GJRGARCH11_CandD_I)
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.10_SPX_CPV_GJRGARCH11_CandD.RData")
#------------------------------------------------------------------------------
# SPXCPV GJRGARCH11SVI's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the GJRGARCH11SVI Model:
SPXCPV_GJRGARCH11SVI_CandD_I = matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_GJRGARCH11SVI_CandD_I_k= matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_GJRGARCH11SVI_CandD_I_t = matrix( - (SPXCPV_GJRGARCH11SVI_mu_hat/SPXCPV_GJRGARCH11SVI_sigma_hat))
for(t in c(1:SPXCPV_roll))
{for (k in c(1:t))
{SPXCPV_GJRGARCH11SVI_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPXCPV_GJRGARCH11SVI_mu_hat[k])/SPXCPV_GJRGARCH11SVI_sigma_hat[k])
<=SPXCPV_GJRGARCH11SVI_CandD_I_t[t+1],
1,0)
SPXCPV_GJRGARCH11SVI_CandD_I[k,t] =
ifelse((is.na(SPXCPV_GJRGARCH11SVI_CandD_I_k[k,t])), NA,
(sum(SPXCPV_GJRGARCH11SVI_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the GJRGARCH11SVI Model:
SPXCPV_GJRGARCH11SVI_CandD=(1:(SPXCPV_roll-1))
for(i in c(1:(SPXCPV_roll-1)))
{SPXCPV_GJRGARCH11SVI_CandD[i] = 1 - ((SPXCPV_GJRGARCH11SVI_CandD_I[i,i])/
length(na.omit(SPXCPV_GJRGARCH11SVI_CandD_I[1:i,i])))}
SPXCPV_GJRGARCH11SVI_CandD_zoo = zoo(SPXCPV_GJRGARCH11SVI_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPXCPV_roll-1))]))
# This zoo object has missing values.
SPXCPV_GJRGARCH11SVI_CandD_zoo_no_nas=zoo(na.omit(SPXCPV_GJRGARCH11SVI_CandD_zoo))
# rm(SPXCPV_GJRGARCH11SVI_CandD_I_t) # Cleaning datasets
# rm(SPXCPV_GJRGARCH11SVI_CandD_I_k)
# rm(SPXCPV_GJRGARCH11SVI_CandD_I)
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.11_SPX_CPV_GJRGARCH11SVI_CandD.RData")
#------------------------------------------------------------------------------
# SPXCPV EGARCH11's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the EGARCH11 Model:
SPXCPV_EGARCH11_CandD_I = matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_EGARCH11_CandD_I_k= matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_EGARCH11_CandD_I_t = matrix( - (SPXCPV_EGARCH11_mu_hat/SPXCPV_EGARCH11_sigma_hat))
for(t in c(1:SPXCPV_roll))
{for (k in c(1:t))
{SPXCPV_EGARCH11_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPXCPV_EGARCH11_mu_hat[k])/SPXCPV_EGARCH11_sigma_hat[k])
<=SPXCPV_EGARCH11_CandD_I_t[t+1],
1,0)
SPXCPV_EGARCH11_CandD_I[k,t] =
ifelse((is.na(SPXCPV_EGARCH11_CandD_I_k[k,t])), NA,
(sum(SPXCPV_EGARCH11_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the EGARCH11 Model:
SPXCPV_EGARCH11_CandD=(1:(SPXCPV_roll-1))
for(i in c(1:(SPXCPV_roll-1)))
{SPXCPV_EGARCH11_CandD[i] = 1 - ((SPXCPV_EGARCH11_CandD_I[i,i])/
length(na.omit(SPXCPV_EGARCH11_CandD_I[1:i,i])))}
SPXCPV_EGARCH11_CandD_zoo = zoo(SPXCPV_EGARCH11_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPXCPV_roll-1))]))
# This zoo object has missing values.
SPXCPV_EGARCH11_CandD_zoo_no_nas=zoo(na.omit(SPXCPV_EGARCH11_CandD_zoo))
# rm(SPXCPV_EGARCH11_CandD_I_t) # Cleaning datasets
# rm(SPXCPV_EGARCH11_CandD_I_k)
# rm(SPXCPV_EGARCH11_CandD_I)
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.12_SPX_CPV_EGARCH11_CandD.RData")
#------------------------------------------------------------------------------
# SPXCPV EGARCH11SVI's C&D Model
#------------------------------------------------------------------------------
# # Create our variables:
# C&D's Indicator function according to the EGARCH11SVI Model:
SPXCPV_EGARCH11SVI_CandD_I = matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_EGARCH11SVI_CandD_I_k= matrix(, nrow=SPXCPV_roll, ncol=SPXCPV_roll)
SPXCPV_EGARCH11SVI_CandD_I_t = matrix( - (SPXCPV_EGARCH11SVI_mu_hat/SPXCPV_EGARCH11SVI_sigma_hat))
for(t in c(1:SPXCPV_roll))
{for (k in c(1:t))
{SPXCPV_EGARCH11SVI_CandD_I_k[k,t] =
ifelse(((SPX_R_matrix[247+k]-SPXCPV_EGARCH11SVI_mu_hat[k])/SPXCPV_EGARCH11SVI_sigma_hat[k])
<=SPXCPV_EGARCH11SVI_CandD_I_t[t+1],
1,0)
SPXCPV_EGARCH11SVI_CandD_I[k,t] =
ifelse((is.na(SPXCPV_EGARCH11SVI_CandD_I_k[k,t])), NA,
(sum(SPXCPV_EGARCH11SVI_CandD_I_k[1:k,t], na.rm = TRUE)))}}
# C&D's model according to the EGARCH11SVI Model:
SPXCPV_EGARCH11SVI_CandD=(1:(SPXCPV_roll-1))
for(i in c(1:(SPXCPV_roll-1)))
{SPXCPV_EGARCH11SVI_CandD[i] = 1 - ((SPXCPV_EGARCH11SVI_CandD_I[i,i])/
length(na.omit(SPXCPV_EGARCH11SVI_CandD_I[1:i,i])))}
SPXCPV_EGARCH11SVI_CandD_zoo = zoo(SPXCPV_EGARCH11SVI_CandD,
as.Date(SPX_Dates[(248+1):(248+(SPXCPV_roll-1))]))
# This zoo object has missing values.
SPXCPV_EGARCH11SVI_CandD_zoo_no_nas=zoo(na.omit(SPXCPV_EGARCH11SVI_CandD_zoo))
# rm(SPXCPV_EGARCH11SVI_CandD_I_t) # Cleaning datasets
# rm(SPXCPV_EGARCH11SVI_CandD_I_k)
# rm(SPXCPV_EGARCH11SVI_CandD_I)
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.13_SPX_CPV_EGARCH11SVI_CandD.RData")
##-----------------------------------------------------------------------------
## compare the predictive performance of each model using SPXCPV data
## (forecast error mean, sd, Brier Scores and Diebold&Mariano statistics)
##-----------------------------------------------------------------------------
# # mean of the probabilistic forecast errors derived from each model
SPXCPV_Observed_pi= ifelse(SPX_R_matrix>0,1,0)
SPXCPV_Observed_pi_zoo = zoo(SPXCPV_Observed_pi, as.Date(SPX_Dates))
SPXCPV_GARCH11_pi_error = SPXCPV_GARCH11_CandD_zoo_no_nas - SPXCPV_Observed_pi_zoo
SPXCPV_GARCH11SVI_pi_error = SPXCPV_GARCH11SVI_CandD_zoo_no_nas - SPXCPV_Observed_pi_zoo
SPXCPV_GJRGARCH11_pi_error = SPXCPV_GJRGARCH11_CandD_zoo_no_nas - SPXCPV_Observed_pi_zoo
SPXCPV_GJRGARCH11SVI_pi_error = SPXCPV_GJRGARCH11SVI_CandD_zoo_no_nas - SPXCPV_Observed_pi_zoo
SPXCPV_EGARCH11_pi_error = SPXCPV_EGARCH11_CandD_zoo_no_nas - SPXCPV_Observed_pi_zoo
SPXCPV_EGARCH11SVI_pi_error = SPXCPV_EGARCH11SVI_CandD_zoo_no_nas - SPXCPV_Observed_pi_zoo
mean(SPXCPV_GARCH11_pi_error)
sd(SPXCPV_GARCH11_pi_error)
mean(SPXCPV_GARCH11SVI_pi_error)
sd(SPXCPV_GARCH11SVI_pi_error)
mean(SPXCPV_GJRGARCH11_pi_error)
sd(SPXCPV_GJRGARCH11_pi_error)
mean(SPXCPV_GJRGARCH11SVI_pi_error)
sd(SPXCPV_GJRGARCH11SVI_pi_error)
mean(SPXCPV_EGARCH11_pi_error)
sd(SPXCPV_EGARCH11_pi_error)
mean(SPXCPV_EGARCH11SVI_pi_error)
sd(SPXCPV_EGARCH11SVI_pi_error)
# # SPXCPV_Brier scores of the probabilistic forecast errors derived from each model
SPXCPV_GARCH11_pi_error_Brier_score =
(1/length(SPXCPV_GARCH11_pi_error))*sum(SPXCPV_GARCH11_pi_error^2)
show(SPXCPV_GARCH11_pi_error_Brier_score)
SPXCPV_GARCH11SVI_pi_error_Brier_score =
(1/length(SPXCPV_GARCH11SVI_pi_error))*sum(SPXCPV_GARCH11SVI_pi_error^2)
show(SPXCPV_GARCH11SVI_pi_error_Brier_score)
SPXCPV_GJRGARCH11_pi_error_Brier_score =
(1/length(SPXCPV_GJRGARCH11_pi_error))*sum(SPXCPV_GJRGARCH11_pi_error^2)
show(SPXCPV_GJRGARCH11_pi_error_Brier_score)
SPXCPV_GJRGARCH11SVI_pi_error_Brier_score =
(1/length(SPXCPV_GJRGARCH11SVI_pi_error))*sum(SPXCPV_GJRGARCH11SVI_pi_error^2)
show(SPXCPV_GJRGARCH11SVI_pi_error_Brier_score)
SPXCPV_EGARCH11_pi_error_Brier_score =
(1/length(SPXCPV_EGARCH11_pi_error))*sum(SPXCPV_EGARCH11_pi_error^2)
show(SPXCPV_EGARCH11_pi_error_Brier_score)
SPXCPV_EGARCH11SVI_pi_error_Brier_score =
(1/length(SPXCPV_EGARCH11SVI_pi_error))*sum(SPXCPV_EGARCH11SVI_pi_error^2)
show(SPXCPV_EGARCH11SVI_pi_error_Brier_score)
# # SPXCPV_Diebold&Mariano statistics
# Here the alternative hypothesis is that method 2 is more accurate than method 1; remember that a small p-value indicates strong evidence against the Null Hypothesis.
SPXCPV_GARCH11_pi_error_dm = SPXCPV_GARCH11_pi_error - (SPXCPV_GARCH11SVI_pi_error*0)
SPXCPV_GARCH11SVI_pi_error_dm = SPXCPV_GARCH11SVI_pi_error - (SPXCPV_GARCH11_pi_error*0)
SPXCPV_GARCH11_dm_test = dm.test(matrix(SPXCPV_GARCH11_pi_error_dm), matrix(SPXCPV_GARCH11SVI_pi_error_dm), alternative = "greater")
SPXCPV_GJRGARCH11_pi_error_dm = SPXCPV_GJRGARCH11_pi_error - (SPXCPV_GJRGARCH11SVI_pi_error*0)
SPXCPV_GJRGARCH11SVI_pi_error_dm = SPXCPV_GJRGARCH11SVI_pi_error - (SPXCPV_GJRGARCH11_pi_error*0)
SPXCPV_GJRGARCH11_dm_test = dm.test(matrix(SPXCPV_GJRGARCH11_pi_error_dm), matrix(SPXCPV_GJRGARCH11SVI_pi_error_dm), alternative = c("greater"))
SPXCPV_EGARCH11_pi_error_dm = SPXCPV_EGARCH11_pi_error - (SPXCPV_EGARCH11SVI_pi_error*0)
SPXCPV_EGARCH11SVI_pi_error_dm = SPXCPV_EGARCH11SVI_pi_error - (SPXCPV_EGARCH11_pi_error*0)
SPXCPV_EGARCH11_dm_test = dm.test(matrix(SPXCPV_EGARCH11_pi_error_dm), matrix(SPXCPV_EGARCH11SVI_pi_error_dm), alternative = c("greater"))
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.14_SPX_CPV_pi_DM.RData")
####----------------------------------------------------------------------------
#### SPXCPV's Financial significance
####----------------------------------------------------------------------------
# indicator of the realised direction of the return on the S&P 500 index
y_d = ifelse(SPX_R_zoo>0,1,0)
# Set the probability threashold at which to invest in the index in our 2d graphs:
for (p in c(0.49, 0.495, 0.5, 0.505, 0.51)){
###----------------------------------------------------------------------------
### SPXCPV Granger and Pesaran (2000)'s framework using the GARCH11 model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPXCPV_GARCH11 = ifelse(SPXCPV_GARCH11_CandD_zoo_no_nas>p,1,0)
y_d_SPXCPV_GARCH11 = y_d - (SPXCPV_GARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GARCH11's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPXCPV_GARCH11),1))==1 , 1, 0)
R_Active_SPXCPV_GARCH11_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPXCPV_GARCH11*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPXCPV_GARCH11*0))))[2:length(y_hat_d_SPXCPV_GARCH11)]
R_Active_SPXCPV_GARCH11_p_cumulated = matrix(,nrow=length(R_Active_SPXCPV_GARCH11_p))
for (i in c(1:length(R_Active_SPXCPV_GARCH11_p)))
{R_Active_SPXCPV_GARCH11_p_cumulated[i] = prod((1+R_Active_SPXCPV_GARCH11_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPXCPV_GARCH11_p))
for (i in c(1:length(R_Active_SPXCPV_GARCH11_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPXCPV_GARCH11_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPXCPV Granger and Pesaran (2000)'s framework using the GARCH11-SVI model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPXCPV_GARCH11SVI = ifelse(SPXCPV_GARCH11SVI_CandD_zoo_no_nas>p,1,0)
y_d_SPXCPV_GARCH11SVI = y_d - (SPXCPV_GARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GARCH11SVI's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPXCPV_GARCH11SVI),1))==1 , 1, 0)
R_Active_SPXCPV_GARCH11SVI_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPXCPV_GARCH11SVI*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPXCPV_GARCH11SVI*0))))[2:length(y_hat_d_SPXCPV_GARCH11SVI)]
R_Active_SPXCPV_GARCH11SVI_p_cumulated = matrix(,nrow=length(R_Active_SPXCPV_GARCH11SVI_p))
for (i in c(1:length(R_Active_SPXCPV_GARCH11SVI_p)))
{R_Active_SPXCPV_GARCH11SVI_p_cumulated[i] = prod((1+R_Active_SPXCPV_GARCH11SVI_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPXCPV_GARCH11SVI_p))
for (i in c(1:length(R_Active_SPXCPV_GARCH11SVI_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPXCPV_GARCH11SVI_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPXCPV Granger and Pesaran (2000)'s framework using the GJRGARCH11 model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPXCPV_GJRGARCH11 = ifelse(SPXCPV_GJRGARCH11_CandD_zoo_no_nas>p,1,0)
y_d_SPXCPV_GJRGARCH11 = y_d - (SPXCPV_GJRGARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GJRGARCH11's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPXCPV_GJRGARCH11),1))==1 , 1, 0)
R_Active_SPXCPV_GJRGARCH11_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPXCPV_GJRGARCH11*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPXCPV_GJRGARCH11*0))))[2:length(y_hat_d_SPXCPV_GJRGARCH11)]
R_Active_SPXCPV_GJRGARCH11_p_cumulated = matrix(,nrow=length(R_Active_SPXCPV_GJRGARCH11_p))
for (i in c(1:length(R_Active_SPXCPV_GJRGARCH11_p)))
{R_Active_SPXCPV_GJRGARCH11_p_cumulated[i] = prod((1+R_Active_SPXCPV_GJRGARCH11_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPXCPV_GJRGARCH11_p))
for (i in c(1:length(R_Active_SPXCPV_GJRGARCH11_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPXCPV_GJRGARCH11_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPXCPV Granger and Pesaran (2000)'s framework using the GJRGARCH11-SVI model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPXCPV_GJRGARCH11SVI = ifelse(SPXCPV_GJRGARCH11SVI_CandD_zoo_no_nas>p,1,0)
y_d_SPXCPV_GJRGARCH11SVI = y_d - (SPXCPV_GJRGARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to GJRGARCH11SVI's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPXCPV_GJRGARCH11SVI),1))==1 , 1, 0)
R_Active_SPXCPV_GJRGARCH11SVI_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPXCPV_GJRGARCH11SVI*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPXCPV_GJRGARCH11SVI*0))))[2:length(y_hat_d_SPXCPV_GJRGARCH11SVI)]
R_Active_SPXCPV_GJRGARCH11SVI_p_cumulated = matrix(,nrow=length(R_Active_SPXCPV_GJRGARCH11SVI_p))
for (i in c(1:length(R_Active_SPXCPV_GJRGARCH11SVI_p)))
{R_Active_SPXCPV_GJRGARCH11SVI_p_cumulated[i] = prod((1+R_Active_SPXCPV_GJRGARCH11SVI_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPXCPV_GJRGARCH11SVI_p))
for (i in c(1:length(R_Active_SPXCPV_GJRGARCH11SVI_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPXCPV_GJRGARCH11SVI_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPXCPV Granger and Pesaran (2000)'s framework using the EGARCH11 model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPXCPV_EGARCH11 = ifelse(SPXCPV_EGARCH11_CandD_zoo_no_nas>p,1,0)
y_d_SPXCPV_EGARCH11 = y_d - (SPXCPV_EGARCH11_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to EGARCH11's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPXCPV_EGARCH11),1))==1 , 1, 0)
R_Active_SPXCPV_EGARCH11_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPXCPV_EGARCH11*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPXCPV_EGARCH11*0))))[2:length(y_hat_d_SPXCPV_EGARCH11)]
R_Active_SPXCPV_EGARCH11_p_cumulated = matrix(,nrow=length(R_Active_SPXCPV_EGARCH11_p))
for (i in c(1:length(R_Active_SPXCPV_EGARCH11_p)))
{R_Active_SPXCPV_EGARCH11_p_cumulated[i] = prod((1+R_Active_SPXCPV_EGARCH11_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPXCPV_EGARCH11_p))
for (i in c(1:length(R_Active_SPXCPV_EGARCH11_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPXCPV_EGARCH11_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPXCPV Granger and Pesaran (2000)'s framework using the EGARCH11-SVI model
###----------------------------------------------------------------------------
# corresponding directional forecast and realised direction
y_hat_d_SPXCPV_EGARCH11SVI = ifelse(SPXCPV_EGARCH11SVI_CandD_zoo_no_nas>p,1,0)
y_d_SPXCPV_EGARCH11SVI = y_d - (SPXCPV_EGARCH11SVI_CandD_zoo_no_nas*0) # This y_d only has values on dates corresponding to EGARCH11SVI's.
# let the portfolio weight attributed to the stock market index be
omega = ifelse( (lead(as.matrix(y_hat_d_SPXCPV_EGARCH11SVI),1))==1 , 1, 0)
R_Active_SPXCPV_EGARCH11SVI_p = ((lag(omega,1) * (SPX_R_zoo-(y_hat_d_SPXCPV_EGARCH11SVI*0)) +
(1-lag(omega,1)) * (US_1MO_r_f_zoo-(y_hat_d_SPXCPV_EGARCH11SVI*0))))[2:length(y_hat_d_SPXCPV_EGARCH11SVI)]
R_Active_SPXCPV_EGARCH11SVI_p_cumulated = matrix(,nrow=length(R_Active_SPXCPV_EGARCH11SVI_p))
for (i in c(1:length(R_Active_SPXCPV_EGARCH11SVI_p)))
{R_Active_SPXCPV_EGARCH11SVI_p_cumulated[i] = prod((1+R_Active_SPXCPV_EGARCH11SVI_p)[1:i])}
R_cumulated = matrix(,nrow=length(R_Active_SPXCPV_EGARCH11SVI_p))
for (i in c(1:length(R_Active_SPXCPV_EGARCH11SVI_p)))
{R_cumulated[i] = prod((1+(SPX_R_zoo+(US_1MO_r_f_zoo - (R_Active_SPXCPV_EGARCH11SVI_p*0))))[1:i])}
###----------------------------------------------------------------------------
### SPXCPV Granger and Pesaran (2000)'s framework 2d Graphs put together
###----------------------------------------------------------------------------
# Plot all 2d polts
plot(zoo(R_Active_SPXCPV_GARCH11_p_cumulated,
as.Date(zoo::index(y_hat_d_SPXCPV_GARCH11[2:length(y_hat_d_SPXCPV_GARCH11)]))),
cex.axis=1, type="l",col="red", xlab='Date', ylim=c(0,3), ylab='SPXCPV strategy gain ($)', main=str_c("SPXCPV Strategy for psi ", p))
lines(zoo(R_cumulated,
as.Date(zoo::index(y_hat_d_SPXCPV_GARCH11[2:length(y_hat_d_SPXCPV_GARCH11)]))),
col="black")
lines(zoo(R_Active_SPXCPV_GARCH11SVI_p_cumulated, as.Date(zoo::index(y_hat_d_SPXCPV_GARCH11SVI[2:length(y_hat_d_SPXCPV_GARCH11SVI)]))),
col="orange")
lines(zoo(R_Active_SPXCPV_GJRGARCH11_p_cumulated,
as.Date(zoo::index(y_hat_d_SPXCPV_GJRGARCH11[2:length(y_hat_d_SPXCPV_GJRGARCH11)]))),
col="blue")
lines(zoo(R_Active_SPXCPV_GJRGARCH11SVI_p_cumulated,
as.Date(zoo::index(y_hat_d_SPXCPV_GJRGARCH11SVI[2:length(y_hat_d_SPXCPV_GJRGARCH11SVI)]))),
col="magenta")
lines(zoo(R_Active_SPXCPV_EGARCH11_p_cumulated,
as.Date(zoo::index(y_hat_d_SPXCPV_EGARCH11[2:length(y_hat_d_SPXCPV_EGARCH11)]))),
col="green")
lines(zoo(R_Active_SPXCPV_EGARCH11SVI_p_cumulated,
as.Date(zoo::index(y_hat_d_SPXCPV_EGARCH11SVI[2:length(y_hat_d_SPXCPV_EGARCH11SVI)]))),
col="purple")
legend(lty=1, cex=1,
"topleft", col=c("black", "red", "orange", "blue", "magenta", "green", "purple"),
legend=c("Buy and hold", "GARCH", "GARCH-SVI", "GJRGARCH", "GJRGARCH-SVI", "EGARCH", "EGARCH-SVI"))
}
save.image("C:/Users/johnukfr/OneDrive/UoE/Disertation/Maths/IDaSRP_077.4.15_SPX_CPV_Finance.RData")
# sink()
####---------------------------------------------------------------------------
#### End???????????????????
####---------------------------------------------------------------------------
#
# # historical averages of positive index returns of up to time
# # t are used to predict the Rising Return:
# R_R_m = mean(Observed_pi_zoo * R)
#
# # historical averages of negative index returns of up to time
# # t are used to predict the Falling Return:
#
#
#
# xi = (1- tau_m) * (1 - tau_f)
#
# w = ifelse(y_hat_d_EGARCH11SVI == 1,
# ifelse(y_d == 1, (R_Active_p), (0)),
# ifelse(y_d == 1, (0), (0)))
#
#
#
# u = function(y_hat_d, y_d)
# {
# ifelse(y_hat_d = 1, R_Active_p_sumed, )
|
|
b4f6013695118af9d13122f2cc3d6c72adf49ab7
|
d5eef5ca98115b14d13345c3e104fe4d9448b721
|
/man/print.SDMfit.Rd
|
6cfe153b07a41dc16c9f27755da43401b32d11ed
|
[] |
no_license
|
giopogg/webSDM
|
b6e3e40fc0ee82f87b2cee3a4c2167555fc3e19a
|
9b011d1dcb58b9e2874f841af4874db752c73fff
|
refs/heads/main
| 2023-04-19T02:10:10.368235
| 2023-03-15T07:23:58
| 2023-03-15T07:23:58
| 359,931,820
| 5
| 1
| null | 2023-03-14T09:38:33
| 2021-04-20T19:39:24
|
R
|
UTF-8
|
R
| false
| true
| 884
|
rd
|
print.SDMfit.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/print.SDMfit.R
\name{print.SDMfit}
\alias{print.SDMfit}
\title{Prints a SDMfit object}
\usage{
\method{print}{SDMfit}(x, ...)
}
\arguments{
\item{x}{A SDMfit object, typically obtained with trophicSDM() and available in the field $model of a trophicSDMfit object}
\item{...}{additional arguments}
}
\value{
Prints a summary of the local SDM
}
\description{
Prints a SDMfit object
}
\examples{
data(Y, X, G)
# define abiotic part of the model
env.formula = "~ X_1 + X_2"
# Run the model with bottom-up control using stan_glm as fitting method and no penalisation
# (set iter = 1000 to obtain reliable results)
m = trophicSDM(Y, X, G, env.formula,
family = binomial(link = "logit"), penal = NULL,
mode = "prey", method = "stan_glm")
m$model$Y1
}
\author{
Giovanni Poggiato
}
|
15f687e7462a61df6dd3a85ee05cbbbcce4955dc
|
e10ccafdbc900072e638285141eab217f241ae2f
|
/man/GeomTimeLineLabel.Rd
|
a0c80ab32c8d9d2f587fce7b01c4775a6e0dfb26
|
[] |
no_license
|
pvisser82/earthquakedata
|
04e545d00a022e4b1fc794c67552b7782e2e1571
|
e04ab6d870acc5f4307a5fc8c8388b196c740d7f
|
refs/heads/master
| 2021-05-06T05:28:06.046277
| 2018-01-02T09:51:34
| 2018-01-02T09:51:34
| 115,095,231
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 617
|
rd
|
GeomTimeLineLabel.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/timeline.R
\docType{data}
\name{GeomTimeLineLabel}
\alias{GeomTimeLineLabel}
\title{GeomTimeLineLabel proto}
\format{An object of class \code{GeomTimeLineLabel} (inherits from \code{Geom}, \code{ggproto}) of length 6.}
\usage{
GeomTimeLineLabel
}
\description{
This geom is responsible for drawing the labels on the timeline. The number of labels are set
using the n_max parameter. The function will retrieve the n_max number of highest magnitudes
using the setup_data function and add the label to those earthquakes.
}
\keyword{datasets}
|
1bc9ff81d4270477ae3f7a898c6b9cb1201ce8d8
|
be232f39144fcd136ce0ba20549124e50d1f495a
|
/1-Exploratory_Data_Analysis.R
|
5ca54fc23c1b24da8260b4c538f0555d33585d7f
|
[] |
no_license
|
davidsalazarv95/Resolve_test
|
f86311e406014e310ba298acc126ce6a424c9176
|
9cb31e331aab7d89b26df084c8b8ddfca1e44fba
|
refs/heads/master
| 2021-07-22T23:11:46.095043
| 2017-11-01T21:10:11
| 2017-11-01T21:10:11
| 109,029,268
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 2,459
|
r
|
1-Exploratory_Data_Analysis.R
|
################ Univariado Valor ################
library(tidyverse)
library(corrplot)
library(hrbrthemes)
library(corrr)
library(viridis)
library(gridExtra)
library(grid)
histograma_valor <- function(datos, binwidth = 50) {
medidas_centrales <- datos %>% select(valor) %>% summarise(mediana = median(valor),
media = mean(valor))
datos %>%
ggplot(aes(x = valor)) +
geom_histogram(binwidth = binwidth, fill = "blue4", alpha = 0.5, color = "black") +
theme_ipsum_rc() +
scale_x_continuous(labels = scales::dollar_format(suffix = "", prefix = "$")) +
geom_vline(xintercept = medidas_centrales[['mediana']], linetype = 4, show.legend = TRUE) +
labs(y = "Conteo",
title = "Histograma de valor de inmueble",
subtitle = paste0("Cada barra representa intervalos de ", binwidth ,". Mediana como línea: ", "$",round(medidas_centrales[['mediana']], 1)))
}
################ Location ################
ubicación <- function(datos) {
datos %>%
ggplot(aes(x = coordx, y = coordy, color = valor, size = valor)) +
geom_point() +
scale_color_viridis() +
theme_ipsum_rc()
}
################ Variación conjunta ################
matriz_correlaciones <- function(datos) {
datos <- datos %>% select(-index) %>% mutate(rio = as.numeric(rio))
corrplot(cor(datos), order = "hclust", tl.col = "black")
}
big_three <- function(datos, num = 3) {
datos <- datos %>% select(-index) %>% mutate(rio = as.numeric(rio))
correlate(datos) %>% stretch() %>% filter(x == "valor") %>% arrange(desc(abs(r))) %>% head(num) %>%
select(y, r) %>% rename(Variable = y, Correlación = r)
}
scatterplots <- function(datos) {
g1 <- datos %>%
ggplot(aes(x = pobreza, y = valor)) +
geom_point(alpha = 0.5) +
geom_smooth(method = "lm", se = FALSE) +
theme_ipsum_rc() +
labs(title = "Valor vs. Pobreza")
g2 <- datos %>%
ggplot(aes(x = cuartos, y = valor)) +
geom_point(alpha = 0.5) +
geom_smooth(method = "lm", se = FALSE) +
theme_ipsum_rc() +
labs(title = "Valor vs. # de Cuartos")
g3 <- datos %>%
ggplot(aes(x = tasaeducativa, y = valor)) +
geom_point(alpha = 0.5) +
geom_smooth(method = "lm") +
theme_ipsum_rc() +
labs(title = "Valor vs. # profesores por alumno")
grid.arrange(g1, g2, g3, ncol = 2, widths = c(1.3, 1.5))
}
|
66488a4355c805221849ff67597c3bda2834700f
|
e6f4e3afd16a7ee5c7a8fb61f7ed697ce88ef4c4
|
/Pro3/R_p3/Subsampling_final_example.R
|
9ed5c50e625af4ca62d0632c0697f24044eba1bb
|
[] |
no_license
|
xl0418/Code
|
01b58d05f7fae1a5fcfec15894ce0ed8c833fd1a
|
75235b913730714d538d6d822a99297da54d3841
|
refs/heads/master
| 2021-06-03T21:10:31.578731
| 2020-11-17T07:50:48
| 2020-11-17T07:50:48
| 136,896,128
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 17,707
|
r
|
Subsampling_final_example.R
|
library(DDD)
library(ggplot2)
library(ggridges)
library(ggthemes)
library(viridis)
library("RColorBrewer")
library(grid)
library(gridExtra)
library(ggtree)
library(ape)
source(paste0(getwd(),'/g_legend.R'))
# source('E:/Code/Pro3/R_p3/barplot3d.R', echo=TRUE)
source('E:/Code/Pro3/R_p3/multi3dbar.R', echo=TRUE)
moviedir = 'E:/Googlebox/Research/Project3/replicate_sim_9sces_results/'
dir = 'E:/Googlebox/Research/Project3/replicate_sim_9sces/'
dir.result = 'E:/Googlebox/Research/Project3/replicate_sim_9sces_results/'
sce.short = c('H','M','L')
scenario = NULL
sce.short.comb.vec = NULL
for(i.letter in sce.short){
for(j.letter in sce.short){
sce.folder = paste0('sce',i.letter,j.letter)
scenario = c(scenario,sce.folder)
sce.short.comb = paste0(i.letter,j.letter)
sce.short.comb.vec = c(sce.short.comb.vec,sce.short.comb)
}
}
interleave <- function(x,y){
lx <- length(x)
ly <- length(y)
n <- max(lx,ly)
as.vector(rbind(rep(x, length.out=n), rep(y, length.out=n)))
}
x_label_fontsize = 12
y_label_fontsize = 12
mu_title_fontsize = 16
mig_title_fontsize = 16
x_title_fontsize = 16
y_title_fontsize = 16
dispersal_title <- rep(c('high dispersal', 'intermediate dispersal', 'low dispersal'), each = 3)
interaction_title <- rep(c('high interaction distance', 'intermediate interaction distance', 'low interaction distance'), 3)
psi_title <- c('0', '0.5', '1', '0.25', '0.75')
phi_title <- c('0', '-2', '-4', '-6', '-8', 0)
subsampling_scale_vec <- rep(c(50,250), 2)
count1 = 1
p = list()
# AWMIPD mode
pdmode = 'exp'
a = 100
rep.sample = 18
plot.combination = rbind(c(1, 3, 3),
c(1, 3, 3),
c(6, 3, 4),
c(6, 3, 4))
# rbind(c(1, 3, 3),
# c(1, 4, 4),
# c(8, 3, 2))
col_labels = NULL
for(plot.comb in c(1:nrow(plot.combination))){
i_n = plot.combination[plot.comb,1]
i = plot.combination[plot.comb,2]
j = plot.combination[plot.comb,3]
subsampling_scale <- subsampling_scale_vec[plot.comb]
comb = paste0(i,j)
scefolder = scenario[i_n]
letter.comb = sce.short.comb.vec[i_n]
sce = scenario[i_n]
print(paste0('i_n = ',i_n,'; comb = ', comb))
multitreefile <- paste0(dir,scefolder,'/results/1e+07/spatialpara1e+07',letter.comb,comb,'/','multitreen',letter.comb,comb,'.tre')
replicate_trees <- read.tree(multitreefile)
single.tree_example <- replicate_trees[[rep.sample]]
rname_example = paste0(dir,scefolder,'/results/1e+07/spatialpara1e+07',letter.comb,comb,'/',letter.comb,'M',i,j,'rep',rep.sample,'.csv')
L.table_example = read.csv(rname_example,header = FALSE)
global.matrix_example = as.matrix(L.table_example)
# sub area grid
sub_local_matrix_example <- global.matrix_example[(167-subsampling_scale/2):(167+subsampling_scale/2),
(167-subsampling_scale/2):(167+subsampling_scale/2)]
species_number_example <- unique(as.vector(sub_local_matrix_example))+1
species_label_example <- paste0('t', species_number_example)
# sub tree
subtree_example <- keep.tip(single.tree_example, species_label_example)
p[[count1]] <- ggtree(subtree_example,layout = "circular") #+xlim(0,age)
count1 = count1+1
# plot SAR
print('Plotting SAR...')
species.area = NULL
comb = paste0(i,j)
for(local.scale in c(1:19,seq(20,subsampling_scale,10))){
mean.data = NULL
print(paste0('i_n = ',i_n,'...','i = ', i, '; j = ',j,'; area = ',local.scale))
for(rep_sar in c(1:100)){
ns.vec=NULL
rname = paste0(dir,scefolder,'/results/1e+07/spatialpara1e+07',letter.comb,comb,'/',letter.comb,'M',i,j,'rep',rep_sar,'.csv')
L.table = read.csv(rname,header = FALSE)
global.matrix = as.matrix(L.table)
sub_local_matrix <- global.matrix[(167-subsampling_scale/2):(167+subsampling_scale/2),
(167-subsampling_scale/2):(167+subsampling_scale/2)]
# calculate abundances in the sub grid
species_number <- unique(as.vector(sub_local_matrix))
submatrix.vec = c(0:(subsampling_scale %/% local.scale - 1))*local.scale+1
for(row.num in submatrix.vec){
local.grid = sub_local_matrix[row.num:(row.num+local.scale-1),row.num:(row.num+local.scale-1)]
local.richness = length(unique(as.vector(as.matrix(local.grid))))
ns.vec = c(ns.vec, local.richness)
}
mean.data = c(mean.data,mean(ns.vec))
}
quantile1 = quantile(mean.data)
species.area = rbind(species.area,c(quantile1,i,j,i_n,local.scale))
}
colnames(species.area) = c('0','25','50','75','100','i','j','i_n','area')
species.area.df <- as.data.frame(species.area)
area = species.area.df$area
df_lineage_sar = rbind(c(rep(1,5),i,j,i_n,1),species.area.df)
df_lineage_sar$`0` = log(df_lineage_sar$`0`)
df_lineage_sar$`25` = log(df_lineage_sar$`25`)
df_lineage_sar$`50` = log(df_lineage_sar$`50`)
df_lineage_sar$`75` = log(df_lineage_sar$`75`)
df_lineage_sar$`100` = log(df_lineage_sar$`100`)
df_lineage_sar$area = log(df_lineage_sar$area)
df_min_max_sar = data.frame(id = "min_max", value = 1, x = c(df_lineage_sar$area,rev(df_lineage_sar$area)), y = c(df_lineage_sar$'0',rev(df_lineage_sar$'100')))
df_0025_sar = data.frame(id = "0025", value = 2, x = c(df_lineage_sar$area,rev(df_lineage_sar$area)), y = c(df_lineage_sar$'25',rev(df_lineage_sar$'75')))
df_mean_sar = data.frame(id = 'mean',y = df_lineage_sar$`50`,x = df_lineage_sar$area)
df_lineage_all_sar = rbind(df_min_max_sar,df_0025_sar)
p[[count1]] <- ggplot(df_mean_sar, aes(x = x, y = y)) +
theme(legend.position="none",
panel.border = element_blank(), panel.grid.major = element_blank(),panel.grid.minor = element_blank(),panel.background = element_blank(),
axis.line.x = element_line(color="black"),axis.line.y = element_line(color="black")
,plot.margin=unit(c(0,0,0,0),"cm"))+
geom_polygon(data = df_min_max_sar, aes( group = id),fill = "gray70", alpha = 0.8)+
geom_polygon(data = df_0025_sar, aes( group = id),fill = "gray27", alpha = 0.8)+
geom_line()+
coord_cartesian(xlim=c(log(1),log(subsampling_scale)),ylim=c(0,log(1000))) +
scale_y_continuous(name = "No. of species",breaks = log(c(1,10,100)),labels = c('1','10','100'))+
scale_x_continuous(name = "Area",breaks = log(c(1,10,100)),labels = c('1','10','100'))+
theme(axis.text.y=element_text(angle=90,size = y_label_fontsize),
axis.text.x=element_text(size = x_label_fontsize))
count1 = count1+1
# SAD
print("Plotting SAD...")
x.breaks = seq(0,17,1)
abund <- NULL
for(rep_sar in c(1:100)){
ns.vec=NULL
rname = paste0(dir,scefolder,'/results/1e+07/spatialpara1e+07',letter.comb,comb,'/',letter.comb,'M',i,j,'rep',rep_sar,'.csv')
L.table = read.csv(rname,header = FALSE)
global.matrix = as.matrix(L.table)
sub_local_matrix <- global.matrix[(167-subsampling_scale/2):(167+subsampling_scale/2),
(167-subsampling_scale/2):(167+subsampling_scale/2)]
# calculate abundances in the sub grid
species_number <- unique(as.vector(sub_local_matrix))
Rs <- NULL
for (spe in species_number) {
Rs <- c(Rs, length(which(as.vector(sub_local_matrix) == spe)))
}
log.Rs = log2(Rs)
freq = hist(as.numeric(log.Rs),plot=FALSE,breaks = x.breaks)
counts = freq$counts
abund = rbind(abund, counts)
}
mean.sim = apply(abund,MARGIN=2,FUN=mean)
sd.sim = sqrt(apply(abund,MARGIN=2,FUN=var))
col.quan = length(mean.sim)
if(col.quan<length(x.breaks)){
mean.sim <- c(mean.sim,matrix(0,1,length(x.breaks)-col.quan))
sd.sim <- c(sd.sim,matrix(0,1,length(x.breaks)-col.quan))
}
abund.df = cbind(mean.sim,sd.sim,c(1:length(x.breaks)))
colnames(abund.df) <- c('mean','sd','species')
abund.df <- as.data.frame(abund.df)
my_labs <- interleave(seq(1,length(x.breaks),2), "")
my_labs = my_labs[1:18]
p[[count1]] <- ggplot(abund.df) +
geom_bar( aes(x=species, y=mean),width = 0.6, stat="identity", fill="red", alpha=0.7) +
geom_errorbar( aes(x=species, ymin=mean-sd, ymax=mean+sd), width=0.4, colour="blue", alpha=0.7, size=1.3)+
geom_line(aes(species, mean),size=0.8,color="blue")+
#theme_gdocs()+ #scale_color_calc()+
scale_x_continuous(name="Abundance (log2)", breaks=seq(1,length(x.breaks),1),labels = my_labs) +
scale_y_continuous(name="Frequency",breaks=seq(0,60,20))+
theme(axis.text.x = element_text(angle = 90,vjust = 0.5),panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
panel.background = element_blank(), axis.line = element_line(colour = "black"),
strip.background = element_blank(),strip.text.x = element_text(size = 12, colour = "black"),
strip.text.y = element_text(size = 12, colour = "black"))
count1 = count1+1
# AWMIPD
print('Plotting AWMIPD...')
# rname = paste0(dir,scefolder,'/results/1e+07/spatialpara1e+07',letter.comb,comb,'/',letter.comb,'M',comb,'rep',rep.sample,'.csv')
# L.table = read.csv(rname,header = FALSE)
Dname_example = paste0(dir,scefolder,'/results/1e+07/spatialpara1e+07',letter.comb,comb,'/',letter.comb,'D',comb,'rep',rep.sample,'.csv')
D.table_example = read.csv(Dname_example,header = FALSE)
Rname_example = paste0(dir,scefolder,'/results/1e+07/spatialpara1e+07',letter.comb,comb,'/',letter.comb,'R',comb,'rep',rep.sample,'.csv')
R.table_example = read.csv(Rname_example,header = FALSE)
R.table_example <- R.table_example[species_number_example]
# global.matrix = as.matrix(L.table)
D.matrix_example = as.matrix(D.table_example) * 2 / 10^7
D.matrix_example = D.matrix_example[species_number_example,species_number_example]
# phylogenetic distance
if (pdmode == 'inv') {
ID = 1 / D.matrix_example
diag(ID) = 1
AD.matrix = sweep(ID, MARGIN = 2, as.matrix(as.numeric(R.table_example)), `*`)
} else if (pdmode == 'exp') {
ID = exp(- a * D.matrix_example)
AD.matrix = sweep(ID, MARGIN = 2, as.matrix(as.numeric(R.table_example)), `*`)
}
total.dvalues = rowSums(AD.matrix) * as.matrix( as.numeric(R.table_example))
D.normalized = total.dvalues/sum(total.dvalues)
D.normalized = (D.normalized-min(D.normalized))/(max(D.normalized)-min(D.normalized))
local.x = c(1:subsampling_scale)
local.y = c(1:subsampling_scale)
distribution.data = expand.grid(X=local.x,Y=local.y)
distribution.data$Z = sub_local_matrix_example[cbind(distribution.data$X,distribution.data$Y)]
distribution.data$D = D.normalized[match((distribution.data$Z+1), species_number_example) ]
p[[count1]] <- ggplot(distribution.data, aes(X, Y, fill= D)) + geom_tile()+
theme(legend.position = '',axis.text = element_blank(),axis.ticks = element_blank(),
panel.background = element_blank())+
xlab("")+ylab("") + scale_fill_gradient2(low="#648FFF",mid = '#DDCC77',
high="#882255",midpoint=0.5)
count1 = count1+1
# LTT plot
print("Plotting LTT...")
subtrees <- list()
tree_index <- 1
for (rep.ltt in c(1:100)) {
single.tree <- replicate_trees[[rep.ltt]]
rname = paste0(dir,scefolder,'/results/1e+07/spatialpara1e+07',letter.comb,comb,'/',letter.comb,'M',i,j,'rep',rep.ltt,'.csv')
M.table = read.csv(rname,header = FALSE)
global.matrix = as.matrix(M.table)
if (length(single.tree$tip.label) == length(unique(as.vector(global.matrix)))) {
# sub area grid
sub_local_matrix <- global.matrix[(167-subsampling_scale/2):(167+subsampling_scale/2),
(167-subsampling_scale/2):(167+subsampling_scale/2)]
species_number <- unique(as.vector(sub_local_matrix))+1
species_label <- paste0('t', species_number)
# sub tree
subtree <- keep.tip(single.tree, species_label)
subtrees[[tree_index]] <- subtree
tree_index <- tree_index+1
} else {
print('Inconsistent diversity...')
}
}
class(subtrees) <- "multiPhylo"
trees <- subtrees
age = 10000000
data = NULL
for (tree_i in 1:length(trees)) {
tes = trees[[tree_i]]
brts= -unname(sort(branching.times(tes),decreasing = T))
data0 = cbind(tree_i,brts,c(2:(length(brts)+1)))
if(max(brts)<0) data0 = rbind(data0,c(tree_i,0,data0[nrow(data0),3]))
data = rbind(data, data0)
}
time = data[order(data[,2]),2]
timeu = unique(time)
data_lineage = timeu
for(tree_i in 1:length(trees)){
tes = trees[[tree_i]]
brts= -unname(sort(branching.times(tes),decreasing = T))
M1 = match(brts,timeu)
M1[1] = 1
M11 = diff(M1)
M13 = length(timeu)-max(M1)+1
M12 = c(M11,M13)
N1 = rep(2:(length(brts)+1),M12)
data_lineage = cbind(data_lineage,N1)
}
x = data_lineage[,1]
z = data_lineage[,2:length(trees)+1]
data_average_z <- apply(z, 1, median)
data_q0.025_z <- apply(z, 1 , quantile, 0.025)
data_q0.25_z <- apply(z, 1, quantile, 0.25)
data_q0.75_z <- apply(z, 1, quantile, 0.75)
data_q0.975_z <- apply(z, 1, quantile, 0.975)
data_lower_z <- apply(z,1,min)
data_upper_z <- apply(z,1,max)
lineage_stat = cbind(x,log(data_average_z),log(data_q0.025_z),log(data_q0.25_z),log(data_q0.75_z),log(data_q0.975_z),log(data_lower_z),log(data_upper_z))
colnames(lineage_stat) = c("time", "median","0.025","0.25","0.75","0.975","min","max")
time = min(lineage_stat[,1])
df_lineage = data.frame(lineage_stat)
df_min_max = data.frame(id = "min_max", value = 1, x = c(df_lineage$time,rev(df_lineage$time)), y = c(df_lineage$min,rev(df_lineage$max)))
df_0025 = data.frame(id = "0025", value = 2, x = c(df_lineage$time,rev(df_lineage$time)), y = c(df_lineage$X0.025,rev(df_lineage$X0.975)))
df_025 = data.frame(id = "025", value = 3, x = c(df_lineage$time,rev(df_lineage$time)), y = c(df_lineage$X0.25,rev(df_lineage$X0.75)))
df_lineage_all = rbind(df_min_max,df_025,df_0025)
p[[count1]] <- ggplot(df_min_max, aes(x = x, y = y)) +
theme(legend.position="none",
panel.border = element_blank(), panel.grid.major = element_blank(),panel.grid.minor = element_blank(),panel.background = element_blank(),
axis.line.x = element_line(color="black"),axis.line.y = element_line(color="black")
,plot.margin=unit(c(0,0,0,0),"cm"))+
geom_polygon(data = df_min_max, aes( group = id),fill = "gray80", alpha = 0.8)+
geom_polygon(data = df_0025, aes( group = id),fill = "gray50", alpha = 0.8)+
geom_polygon(data = df_025, aes( group = id), fill = "gray27", alpha = 0.8)+
coord_cartesian(xlim=c(-age,0),ylim=c(log(2),log(600))) + scale_y_continuous(name="No. of species",breaks = c(log(2),log(10),log(50),log(400)),labels = c(2,10,50,400))+
scale_x_continuous(name="Time",breaks = -rev(seq(0,1e7,1e7/5)),labels = c('10','8','6','4','2','0'))+
theme(axis.text.y=element_text(angle=90,size = y_label_fontsize),axis.text.x=element_text(size = x_label_fontsize))
count1 = count1+1
col_labels <- c(col_labels,
paste0('SPJC ', dispersal_title[i_n], ' &\n', interaction_title[i_n], '\n',
'scale = ',subsampling_scale))
}
m = matrix(1:20,ncol = 4)
row_labels = c('Tree', 'SAR','SAD','SPD','LTT')
phi1 <- textGrob(row_labels[1], gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
phi2 <- textGrob(row_labels[2], gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
phi3 <- textGrob(row_labels[3], gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
phi4 <- textGrob(row_labels[4], gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
phi5 <- textGrob(row_labels[5], gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
psi11 <- textGrob(col_labels[1],
gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
psi12 <- textGrob(col_labels[2],
gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
psi13 <- textGrob(col_labels[3],
gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
psi14 <- textGrob(col_labels[4],
gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
psi21 <- textGrob(bquote(psi == .(psi_title[plot.combination[1, 2]]) ~ phi == 10^.(phi_title[plot.combination[1, 3]])),
gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
psi22 <- textGrob(bquote(psi == .(psi_title[plot.combination[2, 2]]) ~ phi == 10^.(phi_title[plot.combination[2, 3]])),
gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
psi23 <- textGrob(bquote(psi == .(psi_title[plot.combination[3, 2]]) ~ phi == 10^.(phi_title[plot.combination[3, 3]])),
gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
psi24 <- textGrob(bquote(psi == .(psi_title[plot.combination[4, 2]]) ~ phi == 10^.(phi_title[plot.combination[4, 3]])),
gp=gpar(fontsize=mu_title_fontsize, fontface=3L))
row_titles <- arrangeGrob(phi1,phi2,phi3,phi4,phi5,ncol = 1)
column_titles1 <- arrangeGrob(psi11,psi12,psi13,psi14, ncol = 4)
column_titles2 <- arrangeGrob(psi21,psi22,psi23,psi24,ncol = 4)
# label = textGrob("Number of lineages",gp=gpar(fontsize=y_title_fontsize), rot = 90)
g_ltt4 = arrangeGrob(grobs = p, layout_matrix = m)
g_ltt5 = textGrob("area", gp=gpar(fontsize=x_title_fontsize),rot = 0)
g_ltt1 = textGrob("")
g_a = textGrob("(a)")
g_b = textGrob("(b)")
ltt.sce <- grid.arrange(column_titles1,g_ltt1, column_titles2,g_ltt1, g_ltt4,row_titles,ncol = 2,widths = c(24,1),heights = c(2,1,25))
dir_save <- 'E:/Googlebox/Research/Project3/replicate_sim_9sces_results/'
savefilename <- paste0(dir_save,'mixing_example_subscale_a100.png')
ggsave(savefilename,ltt.sce,width = 20,height = 15)
|
7ea35b0a1016904316bebadfef3fd5cbd1eef18c
|
54833a5f2af934ba192600b0c2550f0ce2e4d97f
|
/in-action/apis/analysis.R
|
2cd34a49d3c53e55a5e61f04e57846f28789cd56
|
[
"MIT"
] |
permissive
|
HOXOMInc/programming-skills-for-data-science
|
2e47fb2a445e71bcfe3d7acc2a881fda217e4a2e
|
5aa1e954434674789c5df9518f8f137c3b91358b
|
refs/heads/main
| 2023-06-27T02:03:13.023455
| 2021-07-25T05:00:55
| 2021-07-25T05:00:55
| 319,144,846
| 2
| 4
| null | 2021-01-17T07:56:40
| 2020-12-06T22:28:03
| null |
UTF-8
|
R
| false
| false
| 2,300
|
r
|
analysis.R
|
# APIs in Action: シアトルのキューバレストラン
# 必要なパッケージをロードする
library(httr)
library(jsonlite)
library(dplyr)
library(ggrepel)
# ggmapの開発中のバージョンをインストール・ロードする
library(devtools) # GitHubからパッケージをインストールする
devtools::install_github("dkahle/ggmap", ref = "tidyup")
library(ggmap)
# Google API Keyを登録する
# https://developers.google.com/maps/documentation/geocoding/get-api-key
register_google(key="YOUR_GOOGLE_KEY")
# Yelp APIキーを"api_key.R"からロードする
source("api_key.R") # ロードすることで`yelp_key`変数が使えるようになる
# Yelp Fusion API's Business Searchエンドポイントの検索クエリを作成する
base_uri <- "https://api.yelp.com/v3"
endpoint <- "/businesses/search"
search_uri <- paste0(base_uri, endpoint)
# クエリ変数を代入する
query_params <- list(
term = "restaurant",
categories = "cuban",
location = "Seattle, WA",
sort_by = "rating",
radius = 8000 # ドキュメントに記載あるようにradiusの単位はメートルとなっている
)
# APIキーとクエリ変数を用いてGETリクエストを作成する
response <- GET(
search_uri,
query = query_params,
add_headers(Authorization = paste("bearer", yelp_key))
)
# リクエストの結果をパースする
response_text <- content(response, type = "text")
response_data <- fromJSON(response_text)
# `response_data`変数の内容を調査する
names(response_data) # [1] "businesses" "total" "region"
# flatten()を用いて`response_data$businesses`をデータフレームに変換する
restaurants <- flatten(response_data$businesses)
# データフレームに必要な列を追加する
restaurants <- restaurants %>%
mutate(rank = row_number()) %>% # 行数をランクとする
mutate(name_and_rank = paste0(rank, ". ", name))
# 地図のベースとなるレイヤーを作成する（シアトルのGoogle Maps画像）
base_map <- ggmap(
get_map(
location = c(-122.3321, 47.6062),
zoom = 11,
source = "google")
)
# 地図にラベルを付与する
base_map + geom_label_repel(
data = restaurants,
aes(x = coordinates.longitude, y = coordinates.latitude, label = name_and_rank)
)
|
603b4e282ca403212d04bd1f9473358524542273
|
63e1231faa30a4cea6dd9f25e87c2372383aa2f4
|
/man/L2A.Rd
|
d8b36f47d29d7c36320ae8a050356a14ec4ac7b2
|
[] |
no_license
|
cran/MSEtool
|
35e4f802f1078412d5ebc2efc3149c46fc6d13a5
|
6b060d381adf2007becf5605bc295cca62f26770
|
refs/heads/master
| 2023-08-03T06:51:58.080968
| 2023-07-19T22:10:23
| 2023-07-20T01:47:18
| 145,912,213
| 1
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 661
|
rd
|
L2A.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/Misc_Exported.R
\name{L2A}
\alias{L2A}
\title{Length to age conversion}
\usage{
L2A(t0c, Linfc, Kc, Len, maxage, ploty = F)
}
\arguments{
\item{t0c}{Theoretical age at length zero}
\item{Linfc}{Maximum length}
\item{Kc}{Maximum growth rate}
\item{Len}{Length}
\item{maxage}{Maximum age}
\item{ploty}{Should a plot be included}
}
\value{
An age (vector of ages, matrix of ages) corresponding with Len
}
\description{
Simple deterministic length to age conversion given inverse von Bertalanffy
growth.
}
\author{
T. Carruthers
}
\keyword{internal}
|
4f678fec2c41c02a12223a5edc4836843ffa880c
|
15f26ad8f0fef7e64ab39a39fcfcf501ae15c5f1
|
/man/zipf_race.Rd
|
5340181924bedcc822f0696f7fff4eda1ab04098
|
[] |
no_license
|
seqva/RcappeR
|
56120272f9b39287e579134346e268800fd78bf6
|
e84bdf0b386898fc4b88850cb298efaa442cdca0
|
refs/heads/master
| 2020-07-02T15:13:55.485347
| 2016-02-26T00:02:30
| 2016-02-26T00:02:30
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 1,437
|
rd
|
zipf_race.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/zipf_race.R
\name{zipf_race}
\alias{zipf_race}
\title{Handicap a race using one race}
\source{
Article by Simon Rowlands explaining use of Zipf's Law:
\url{https://betting.betfair.com/horse-racing/bloggers/simon-rowlands/simon-rowlands-on-handicapping-060710.html}
}
\usage{
zipf_race(race, btn_var, race_2, rating = NULL)
}
\arguments{
\item{race}{dataframe of race to handicap}
\item{btn_var}{name of variable which contains the margins between
the horses}
\item{race_2}{dataframe of a race to be used to handicap
\strong{race}}
\item{rating}{name of ratings variable (if applicable) in
\strong{race_2}}
}
\description{
Assess the performance of a winner (ie handicap) of a race using race
standardisation; which uses the performances of runners in a different, but
similar, race for this assessment. It is called by \link{zipf_hcp} and
\link{zipf_init}.
}
\details{
The method of race standardisation used was first explained by Simon
Rowlands and uses Zipfs Law. The difference at the weights, from the race
to be handicapped, is applied to the second race, creating a vector of
possible ratings the victor could have achieved. This is where Zipfs Law
plays its part, as depending on the finishing position, different weights
are assigned to each of the ratings in the vector, placing more significance
on horses towards the front of the field.
}
|
3fc057d66e362395b0c16524e5dd5b34ef1fa2ee
|
eec115235405a54b642d3286863dcca14786c9e8
|
/experiments/experiment_variance_pvalue_plotter.R
|
62f1cd1ef596df03b50c504aea082e213e20eb55
|
[] |
no_license
|
linnykos/selectiveModel
|
5f5c95f84c006dc8bd83fdc2032ea251a9f4d3e6
|
75e42567a9a4fe1daf4d80a54cc15e47e549f60d
|
refs/heads/master
| 2023-01-21T18:24:22.665232
| 2020-12-01T16:45:43
| 2020-12-01T16:45:43
| 110,089,332
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 550
|
r
|
experiment_variance_pvalue_plotter.R
|
png("../figures/pvalue_stability.png", heigh = 2000, width = 2500, units = "px",
res = 300)
par(mfrow = c(2,2))
hist(res_vec_known_0, breaks = seq(0, 1, length.out = 50), col = "gray",
xlab = "P-value", main = "Delta = 0 (Null), Known sigma")
hist(res_vec_unknown_0, breaks = seq(0, 1, length.out = 50), col = "gray",
xlab = "P-value", main = "Delta = 0 (Null), Unknown sigma")
hist(res_vec_known_signal, breaks = seq(0, 1, length.out = 50), col = "gray",
xlab = "P-value", main = "Delta = 1 (Signal), Known sigma")
graphics.off()
|
ea6a9546894cf4023fe51412c984a1663c14c325
|
2a6dc4ea444f5522291923ef172865314c4bc9e8
|
/R/sgdWt_convexLinCom.R
|
f01dace307c7cc895353b5db58bb578fc93a64f7
|
[] |
no_license
|
benkeser/onlinesl
|
e482e8606268a546011d6bc01019bb5e51c30ad3
|
acf1935523cf5bdf4c0408f55f7930e74bc7f384
|
refs/heads/master
| 2020-12-26T04:27:31.876193
| 2016-09-22T15:41:39
| 2016-09-22T15:41:39
| 66,727,503
| 3
| 1
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,228
|
r
|
sgdWt_convexLinCom.R
|
#' Perform first order stochastic gradient descent update of super learner
#' weights
#'
#' This function performs a single step of gradient descent on the weight
#' vector for the super learner weights and projects the resulting vector onto
#' the L1-simplex via the internal function .projToL1Simp. The function returns
#' the updated weight vector.
#'
#' @param Y The outcome at iteration t
#' @param slFit.t A named list with a component named alpha.t that contains the
#' 1-column matrix of current estimate of the super learner weights
#' @param p.t The predictions from the various online algorithms at time t
#' @param tplus1 The iteration of the online algorithm
#' @param stepSize The size of the step to take in the direction of the
#' gradient. If \code{stepSize=NULL} (default) the function uses
#' \code{1/tplus1}.
#'
#' @return alpha A matrix of updated weights.
#'
#' @export
sgdWt_convexLinCom <- function(Y,slFit.t,p.t,tplus1,stepSize=NULL){
if(is.null(stepSize)){
stepSize <- 1/tplus1
}
grad <- - t(p.t) %*% (Y - p.t%*%slFit.t$alpha)
cwt <- slFit.t$alpha - stepSize * grad
wt.tplus1 <- onlinesl:::.projToL1Simp(cwt)
list(alpha=wt.tplus1)
}
|
bebc5f4c8b628eb1df2a6ed9317a3a70edfb8f75
|
d966e9fe4e0e667b20cad92c1cd21a55ccf401a9
|
/R/TitanicAnalysis01_lowlevel.R
|
7153869eccee011ae6bffcff697c456e0f92c140
|
[] |
no_license
|
darsa881r/titanic
|
0efbb92d408958de5860e431206319403e87e71a
|
f519ab108bf7eede6c9f5126cc33d5228062ae4b
|
refs/heads/master
| 2022-11-26T17:03:12.945916
| 2020-08-10T20:52:28
| 2020-08-10T20:52:28
| 286,527,686
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 30,361
|
r
|
TitanicAnalysis01_lowlevel.R
|
# Loading default Libraries
library("ggplot2")
library("datasets")
library("graphics")
library("grDevices")
library("methods")
library("stats")
library("utils")
# Loading data
train <- read.csv("/Users/sabbirhassan/Dropbox/ML_stuff/titanic/train.csv", header = TRUE)
test <- read.csv("/Users/sabbirhassan/Dropbox/ML_stuff/titanic/test.csv", header = TRUE)
#train <- read.csv("D:\\Dropbox\\ML_stuff\\titanic\\train.csv", header = TRUE)
#test <- read.csv("D:\\Dropbox\\ML_stuff\\titanic\\test.csv", header = TRUE)
# Adding Survived empty valued column in test set
test.survived <- data.frame(Survived = rep("None",nrow(test)),test[,])
# Combining the two data sets
data.combined <- rbind(train, test.survived)
#look at the datatypes
str(data.combined)
# Pclass seems like int but actually its an factor (or categorical) data
# Survived also seems char type but its supposed to be a factortype
data.combined$Pclass <- as.factor(data.combined$Pclass)
data.combined$Survived <- as.factor(data.combined$Survived)
# Take a look at survival frequencies
table(data.combined$Survived)
table(data.combined$Pclass)
# But instead of tabular representation we visualize for more insight
train$Pclass <- as.factor(train$Pclass)
train$Survived <- as.factor(train$Survived)
test$Pclass <- as.factor(test$Pclass)
ggplot(train, aes(x = Pclass, fill = factor(Survived))) +
geom_bar(width = 0.5) +
xlab("Pclass") +
ylab("Total Count") +
labs(fill = "Survived")
# This shows the society class does play a role in survival
head(as.character((train$Name)))
length(unique(as.character(data.combined$Name)))
# Lets find the duplicate names
dup.names <- as.character(data.combined[which(duplicated(as.character(data.combined$Name))), "Name"])
# Now lets pullout this name and check their records
data.combined[which(data.combined$Name %in% dup.names),]
# So we can decide they are different people
# Now we see Miss , Mr, Mrs are inside the Names, So lets extract them out
library(stringr)
misses <- data.combined[which(str_detect(data.combined$Name, "Miss.")),]
summary(misses)
mrs <- data.combined[which(str_detect(data.combined$Name, "Mrs.")),]
males <- data.combined[which(train$Sex == "male"),]
# Do something New with function
# extract title using a function
extractTitle <- function(Name) {
name <- as.character(Name)
if (length(grep("Miss.", name))>0) {
return ("Miss.")
} else if (length(grep("Master.", name))>0) {
return("Master.")
} else if (length(grep("Mrs.", name))>0) {
return("Mrs.")
} else if (length(grep("Mr.", name))>0) {
return("Mr.")
} else {
return("Other")
}
}
titles <- NULL
for (i in 1:nrow(data.combined)) {
titles <- c(titles, extractTitle(data.combined[i,"Name"]))
}
data.combined$title <- as.factor(titles)
ggplot(data.combined[1:891,], aes(x = title, fill= factor(Survived))) +
geom_bar(width = 0.5) +
facet_wrap(~Pclass) +
ggtitle("Pclass") +
xlab("Title") +
ylab("total count")+
labs(fill = "Survived")
# look at sex
table(data.combined$Sex)
ggplot(data.combined[1:891,], aes(x = Sex, fill= factor(Survived))) +
geom_bar(width = 0.5) +
facet_wrap(~Pclass) +
ggtitle("Pclass") +
xlab("Sex") +
ylab("Total Count")+
labs(fill = "Survived")
# look at Age
summary(data.combined$Age)
summary(data.combined[1:891,"Age"])
# we see a lot of missing values; so how to deal with them?
# what are the imputation methods? Gradient boosting trees: xgboost, lightgbm ???
ggplot(data.combined[1:891,], aes(x = Age, fill= factor(Survived))) +
geom_bar(width = 10) +
facet_wrap(~Sex+Pclass) +
ggtitle("Pclass") +
xlab("Age") +
ylab("Total Count") +
labs(fill = "Survived")
# validate that master is a good proxy of male choldren
boys <- data.combined[which(data.combined$title == "Master."),]
summary(boys$Age)
girls <- data.combined[which(data.combined$title == "Miss."),]
summary(girls$Age)
adultmales <- data.combined[which(data.combined$title == "Mr."),]
summary(adultmales$Age)
adultfemales <- data.combined[which(data.combined$title == "Mrs."),]
summary(adultfemales$Age)
ggplot(misses[misses$Survived != "None",], aes(x = Age, fill= factor(Survived))) +
geom_bar(width = 5) +
facet_wrap(~Pclass) +
ggtitle("Pclass") +
xlab("Age") +
ylab("Total Count")
labs(fill = "Survived")
# looking closer into female misses
misses.alone <- misses[which(misses$SibSp == 0 & misses$Parch == 0),]
summary(misses.alone$Age)
length(which(misses.alone$Age <= 14.5))
# this insight is helpful for feature engineering
# look look into Sibsp variable
summary(data.combined$SibSp) # can we treat this as a categorical variable? as its
#finite value, and make a dropdown list to select or sth.
length(unique(data.combined$SibSp))
# only 7, so we can turn it into factors
ggplot(data.combined[1:891,], aes(x = SibSp, fill= factor(Survived))) +
geom_bar(width = 0.5) +
facet_wrap(~Pclass + title) +
ggtitle("Pclass, title") +
xlab("SibSp") +
ylab("Total Count") +
ylim(0,300) +
labs(fill = "Survived")
ggplot(data.combined[1:891,], aes(x = as.factor(Parch), fill= as.factor(Survived))) +
geom_bar(width = 0.5) +
facet_wrap(~Pclass + title) +
ggtitle("Pclass, title") +
xlab("Parch") +
ylab("Total Count") +
ylim(0,300) +
labs(fill = "Survived")
#Create a family size feature
temp.SibSp <- c(train$SibSp, test$SibSp)
temp.Parch <- c(train$Parch, test$Parch)
data.combined$family <- as.factor(temp.Parch+temp.SibSp+1)
#Now lets look into this new feature
ggplot(data.combined[1:891,], aes(x = as.factor(family), fill= as.factor(Survived))) +
geom_bar(width = 0.5) +
facet_wrap(~Pclass + title) +
ggtitle("Pclass, title") +
xlab("Family Size") +
ylab("Total Count") +
ylim(0,300) +
labs(fill = "Survived")
# Looking into tickets and fares
str(data.combined$Ticket)
data.combined$Ticket <- as.character(data.combined$Ticket)
# lets start looking at just the frst character for each
Ticket.first.char <- ifelse(data.combined$Ticket == ""," ", substr(data.combined$Ticket,1,1))
unique(Ticket.first.char) #we can make it a factor
data.combined$Ticket.first.char <- as.factor(Ticket.first.char)
ggplot(data.combined[1:891,], aes(x = as.factor(Ticket.first.char), fill= as.factor(Survived))) +
geom_bar(width = 0.5) +
facet_wrap(~Pclass + title) +
ggtitle("Pclass, title") +
xlab("Ticket First Char") +
ylab("Total Count") +
ylim(0,200) +
labs(fill = "Survived")
ggplot(data.combined[1:891,], aes(x = as.factor(Ticket.first.char), fill= as.factor(Survived))) +
geom_bar(width = 0.5) +
#facet_wrap(~Pclass + title) +
#ggtitle("Pclass, title") +
xlab("Ticket First Char") +
ylab("Total Count") +
ylim(0,200) +
labs(fill = "Survived")
# there probably very strng correlation betn class and fare price. it's obvious, but lets take a look at it
str(data.combined$Fare)
summary(data.combined$Fare)
length(unique(data.combined$Fare))
ggplot(data.combined[1:891,], aes(x = Fare, fill= as.factor(Survived))) +
geom_bar(width = 5, position = "stack") +
#facet_wrap(~Pclass + title) +
ggtitle("Fare Distribution") +
xlab("Fare") +
ylab("Total Count") +
ylim(0,50) +
labs(fill = "Survived")
#probably fare doesnt affet much?
#analysis in caboin variable
str(data.combined$Cabin)
length(unique(as.character(data.combined$Cabin)))
# random forest is looking to be useful here, but default factor levels RF can use is 32
summary(data.combined$Cabin)
# seems that some blank maybe there
data.combined$Cabin[1:100]
#replace empty cabin with U
data.combined$Cabin <- as.character(data.combined$Cabin)
data.combined[(as.character(data.combined$Cabin) == ""),"Cabin"] <-"U"
#from cabin names we can see that the frst char might be importatn
Cabin.first.char <- substr(data.combined$Cabin,1,1)
str(Cabin.first.char)
unique(Cabin.first.char) #we can make it a factor
data.combined$Cabin.first.char <- as.factor(Cabin.first.char)
ggplot(data.combined[1:891,], aes(x = as.factor(Cabin.first.char), fill= as.factor(Survived))) +
geom_bar(width = 0.5) +
facet_wrap(~Pclass + title) +
ggtitle("Pclass, title") +
xlab("Cabin First Char") +
ylab("Total Count") +
ylim(0,700) +
labs(fill = "Survived")
ggplot(data.combined[1:891,], aes(x = as.factor(Cabin.first.char), fill= as.factor(Survived))) +
geom_bar(width = 0.5) +
#facet_wrap(~Pclass + title) +
#ggtitle("Pclass, title") +
xlab("Cabin First Char") +
ylab("Total Count") +
ylim(0,700) +
labs(fill = "Survived")
#seemes like strong correlationwit Pclass, so no new info, not thathelpful !!
# what about ppl with multiple cabins
data.combined$Cabin.multiple <- as.factor(ifelse(str_detect(data.combined$Cabin, " "),"Y","N"))
ggplot(data.combined[1:891,], aes(x = as.factor(Cabin.multiple), fill= as.factor(Survived))) +
geom_bar(width = 0.5) +
facet_wrap(~Pclass + title) +
ggtitle("Pclass, title") +
xlab("Cabin multiple") +
ylab("Total Count") +
ylim(0,200) +
labs(fill = "Survived")
#seems not that important , even though RF algorithm will tell us later
# checking Embraked column
ggplot(data.combined[1:891,], aes(x = as.factor(Embarked), fill= as.factor(Survived))) +
geom_bar(width = 0.5) +
facet_wrap(~Pclass + title, nrow = 3, ncol = 5) +
ggtitle("Pclass, title") +
xlab("Embarked") +
ylab("Total Count") +
ylim(0,100) +
labs(fill = "Survived")
# So exploratory analysis tells us that some variables asre important and some can be replaced by our
#derived features, but what evidence can we use to rule a column out?
#Now how we can get the idea of feature importance? some models do it implicitly : RF
#sometimes we need to do PCA to know importance.
#Lets Do Random forest and then upgrade to Xgboost later . (also show that Logistic is not good)
#what about neural network for categorical data
#install.packages("randomForest")
library(randomForest)
#tarin a random forest with default parameters
#lets check all the vars
# rf.train.1 <- data.combined[1:891,c("Pclass","Sex","SibSp","Parch","Fare","Embarked","title","family","Ticket.first.char","Cabin.first.char","Cabin.multiple")] #taking just two important cols that we think
# #we drop Age due to missing values
# rf.label <- as.factor(train$Survived)
#
# set.seed(1) #seting seed to verify a random run
# rf.1 <- randomForest(x=rf.train.1, y=rf.label, importance = TRUE, ntree = 1000)
#
# rf.1 #just this output gives some info esp the confusion matrix and oob (out of the bag)
# importance(rf.1) #getting importance values
# varImpPlot(rf.1)
#taking the frst 4 variables
rf.train.2 <- data.combined[1:891,c("title","family","Ticket.first.char","Cabin.first.char")] #taking just two important cols that we think
#we drop Age due to missing values
rf.label <- as.factor(train$Survived)
set.seed(1) #seting seed to verify a random run
rf.2 <- randomForest(x=rf.train.2, y=rf.label, importance = TRUE, ntree = 1000)
rf.2 #just this output gives some info esp the confusion matrix and oob (out of the bag)
importance(rf.2) #getting importance values
varImpPlot(rf.2)
#Making predictions
# Ommiting NA and predicting
test.submit.df <- data.combined[892:1309, c("title","family","Ticket.first.char","Cabin.first.char")]
rf.2.pred <- predict(rf.2, test.submit.df)
submit.df <- data.frame(PassengerID = rep(892:1309),Survived = rf.2.pred)
write.csv(submit.df,file = "/Users/sabbirhassan/Dropbox/ML_stuff/titanic/RF_SUB_20190205_1.csv", row.names = FALSE)
#installed.packages(caret)
library(caret) # good for Crossvalidation process; but its really expensive in terms of computaitons
#install.packages(doSNOW)
library(doSNOW)
# default cross validation settings is 10-fold cross validation repeated 10 times is standard
# caret works parallely and uses cores of the PC.
#doSNOW works on both MAc and Windows using parallelizinf the work for caret
# so we need 10 * 10 = 100 fols that are random.
#
#
# set.seed(5)
# cv.10.folds <- createMultiFolds(rf.label, k = 10, times = 10) #creates bunch of indexes
#
# #check stratification , caret can do it for you.
# #its usefull for small amount of imbalancing
#
# table(rf.label)
# 342/549
#
# table(rf.label[cv.10.folds[[32]]])
# 308/494
#
# #seems like the ratio is okay
# library("e1071")
# cntrl.1 <- trainControl(method = "repeatedcv", number = 10, repeats = 10,
# index = cv.10.folds)
# # setting up doSNOW package for multi-core training; doMC only works on Linux
# cl <- makeCluster(4, type = "SOCK")
# registerDoSNOW(cl)
#
# set.seed(10)
# rf.2.cv.1 <- train(x = rf.train.2, y = rf.label, method = "rf", tuneLength = 3,
# ntree = 1000, trControl = cntrl.1)
#
# #Shut down cluster
# stopCluster(cl)
#
# #checkout results
#
# rf.2.cv.1
#
# #maybe 10 folds overfit... what happens if we do 5 fold?
# set.seed(15)
# cv.5.folds <- createMultiFolds(rf.label, k = 5, times = 10) #creates bunch of indexes
#
# #check stratification , caret can do it for you.
# #its usefull for small amount of imbalancing
#
#
# cntrl.2 <- trainControl(method = "repeatedcv", number = 10, repeats = 10,
# index = cv.5.folds)
# # setting up doSNOW package for multi-core training; doMC only works on Linux
# cl <- makeCluster(4, type = "SOCK")
# registerDoSNOW(cl)
#
# set.seed(20)
# rf.2.cv.2 <- train(x = rf.train.2, y = rf.label, method = "rf", tuneLength = 3,
# ntree = 1000, trControl = cntrl.2)
#
# #Shut down cluster
# stopCluster(cl)
#
# #checkout results
#
# rf.2.cv.2
#
set.seed(25)
cv.3.folds <- createMultiFolds(rf.label, k = 3, times = 10) #creates bunch of indexes
#check stratification , caret can do it for you.
#its usefull for small amount of imbalancing
cntrl.3 <- trainControl(method = "repeatedcv", number = 10, repeats = 10,
index = cv.3.folds)
# setting up doSNOW package for multi-core training; doMC only works on Linux
cl <- makeCluster(4, type = "SOCK")
registerDoSNOW(cl)
set.seed(30)
rf.2.cv.3 <- train(x = rf.train.2, y = rf.label, method = "rf", tuneLength = 3,
ntree = 64, trControl = cntrl.3) #also reducing the trees
#Shut down cluster
stopCluster(cl)
#checkout results
rf.2.cv.3
#so this 3-fold maybe more generalized and better.
test.submit.df <- data.combined[892:1309, c("title","family","Ticket.first.char","Cabin.first.char")]
rf.2.cv.3.pred <- predict(rf.2.cv.3, test.submit.df)
submit.df <- data.frame(PassengerID = rep(892:1309),Survived = rf.2.cv.3.pred)
write.csv(submit.df,file = "/Users/sabbirhassan/Dropbox/ML_stuff/titanic/RF_SUB_20190205_2.csv", row.names = FALSE)
# a small improvement in Kaggle 0.6 %
# the link has a basic tutorial using caret
# https://www.analyticsvidhya.com/blog/2016/12/practical-guide-to-implement-machine-learning-with-caret-package-in-r-with-practice-problem/
# Let's look into the details of a decision tree to understand
# install.packages("rpart")
# install.packages("rpart.plot")
# library("rpart")
# library("rpart.plot")
# LEts update the titles and look at it more closely
# Parse out last name and title
data.combined[1:25, "Name"]
name.splits <- str_split(data.combined$Name, ",")
name.splits[1]
last.names <- sapply(name.splits, "[", 1) # paralell and vectorize applies a function on all elements
last.names[1:10] # "[" is the indexing element
# Add last names to dataframe in case we find it useful later
data.combined$last.name <- last.names
# Now for titles
name.splits <- str_split(sapply(name.splits, "[", 2), " ")
titles <- sapply(name.splits, "[", 2)
unique(titles)
# What's up with a title of 'the'?
data.combined[which(titles == "the"),]
# Re-map titles to be more exact
titles[titles %in% c("Dona.", "the")] <- "Lady."
titles[titles %in% c("Ms.", "Mlle.")] <- "Miss."
titles[titles == "Mme."] <- "Mrs."
titles[titles %in% c("Jonkheer.", "Don.")] <- "Sir."
titles[titles %in% c("Col.", "Capt.", "Major.")] <- "Officer"
table(titles)
# Make title a factor
data.combined$New.title <- as.factor(titles)
# Visualize new version of title
ggplot(data.combined[1:891,], aes(x = New.title, fill = Survived)) +
geom_bar() +
facet_wrap(~ Pclass) +
ggtitle("Surival Rates for new.title by pclass")
# Collapse titles based on visual analysis
indexes <- which(data.combined$New.title == "Lady.")
data.combined$New.title[indexes] <- "Mrs."
indexes <- which(data.combined$New.title == "Dr." |
data.combined$New.title == "Rev." |
data.combined$New.title == "Sir." |
data.combined$New.title == "Officer")
data.combined$New.title[indexes] <- "Mr."
# Visualize
ggplot(data.combined[1:891,], aes(x = New.title, fill = Survived)) +
geom_bar() +
facet_wrap(~ Pclass) +
ggtitle("Surival Rates for Collapsed new.title by pclass")
# Grab features
features <- c("Pclass", "New.title","family","Ticket.first.char","Cabin.first.char")
# Dive in on 1st class Mr."
indexes.first.mr <- which(data.combined$New.title == "Mr." & data.combined$Pclass == "1")
first.mr.df <- data.combined[indexes.first.mr, ]
summary(first.mr.df)
# One female? so this Dr is a woman
first.mr.df[first.mr.df$Sex == "female",]
# Update new.title feature
indexes <- which(data.combined$New.title == "Mr." &
data.combined$Sex == "female")
data.combined$New.title[indexes] <- "Mrs."
# Any other gender slip ups?
length(which(data.combined$Sex == "female" &
(data.combined$New.title == "Master." |
data.combined$New.title == "Mr.")))
# Refresh data frame
indexes.first.mr <- which(data.combined$New.title == "Mr." & data.combined$Pclass == "1")
first.mr.df <- data.combined[indexes.first.mr, ]
# Let's look at surviving 1st class "Mr."
summary(first.mr.df[first.mr.df$Ssurvived == "1",])
View(first.mr.df[first.mr.df$Survived == "1",])
# Take a look at some of the high fares
# Visualize survival rates for 1st class "Mr." by fare
ggplot(first.mr.df, aes(x = Fare, fill = Survived)) +
geom_density(alpha = 0.5) +
ggtitle("1st Class 'Mr.' Survival Rates by fare")
# Engineer features based on all the passengers with the same ticket
# getting per person avg fare. instead of whole pclass fare
ticket.party.size <- rep(0, nrow(data.combined))
avg.fare <- rep(0.0, nrow(data.combined))
tickets <- unique(data.combined$Ticket)
for (i in 1:length(tickets)) {
current.ticket <- tickets[i]
party.indexes <- which(data.combined$Ticket == current.ticket)
current.avg.fare <- data.combined[party.indexes[1], "Fare"] / length(party.indexes)
for (k in 1:length(party.indexes)) {
ticket.party.size[party.indexes[k]] <- length(party.indexes)
avg.fare[party.indexes[k]] <- current.avg.fare
}
}
data.combined$ticket.party.size <- ticket.party.size
data.combined$avg.fare <- avg.fare
# Refresh 1st class "Mr." dataframe
first.mr.df <- data.combined[indexes.first.mr, ]
summary(first.mr.df)
# Visualize new features
ggplot(first.mr.df[first.mr.df$Survived != "None",], aes(x = ticket.party.size, fill = Survived)) +
geom_density(alpha = 0.5) +
ggtitle("Survival Rates 1st Class 'Mr.' by ticket.party.size")
ggplot(first.mr.df[first.mr.df$Survived != "None",], aes(x = avg.fare, fill = Survived)) +
geom_density(alpha = 0.5) +
ggtitle("Survival Rates 1st Class 'Mr.' by avg.fare")
# Hypothesis - ticket.party.size is highly correlated with avg.fare
summary(data.combined$avg.fare)
# One missing value, take a look
data.combined[is.na(data.combined$avg.fare), ]
# Get records for similar passengers and summarize avg.fares
indexes <- with(data.combined, which(Pclass == "3" & title == "Mr." & family == 1 &
Ticket != "3701"))
similar.na.passengers <- data.combined[indexes,]
summary(similar.na.passengers$avg.fare)
# Use median since close to mean and a little higher than mean
data.combined[is.na(avg.fare), "avg.fare"] <- 7.840
#summary(similar.na.passengers$Age) will give you NA in age. and its a lot.
#it requires a lot of imputation. so we ignore it now
# normalizing data is very important to esp for models like support vector machine and others
# Leverage caret's preProcess function to normalize data using z-score
preproc.data.combined <- data.combined[, c("ticket.party.size", "avg.fare")]
preProc <- preProcess(preproc.data.combined, method = c("center", "scale"))
postproc.data.combined <- predict(preProc, preproc.data.combined)
# Hypothesis refuted for all data; seesm like uncorrelated so we can use these as features
cor(postproc.data.combined$ticket.party.size, postproc.data.combined$avg.fare)
# How about for just 1st class all-up?
indexes <- which(data.combined$Pclass == "1")
cor(postproc.data.combined$ticket.party.size[indexes],
postproc.data.combined$avg.fare[indexes])
# Hypothesis refuted again
# OK, let's see if our feature engineering has made any difference
#including previous plus the new and checking the importance
features <- c("Pclass", "New.title","family","Ticket.first.char","Cabin.first.char", "ticket.party.size", "avg.fare")
rf.train.3 <- data.combined[1:891, features]
rf.label <- as.factor(train$Survived)
set.seed(40) #seting seed to verify a random run
rf.3 <- randomForest(x=rf.train.3, y=rf.label, importance = TRUE, ntree = 1000)
rf.3 #just this output gives some info esp the confusion matrix and oob (out of the bag)
importance(rf.3) #getting importance values
varImpPlot(rf.3)
test.submit.df <- data.combined[892:1309, features]
rf.3.pred <- predict(rf.3, test.submit.df)
submit.df <- data.frame(PassengerID = rep(892:1309),Survived = rf.3.pred)
write.csv(submit.df,file = "/Users/sabbirhassan/Dropbox/ML_stuff/titanic/RF_SUB_20190207_03.csv", row.names = FALSE)
#lets reduce the features and check
features <- c("Pclass", "New.title","family","ticket.party.size", "avg.fare")
rf.train.4 <- data.combined[1:891, features]
rf.label <- as.factor(train$Survived)
set.seed(45) #seting seed to verify a random run
rf.4 <- randomForest(x=rf.train.4, y=rf.label, importance = TRUE, ntree = 1000)
rf.4 #just this output gives some info esp the confusion matrix and oob (out of the bag)
importance(rf.4) #getting importance values
varImpPlot(rf.4)
test.submit.df <- data.combined[892:1309, features]
rf.4.pred <- predict(rf.4, test.submit.df)
submit.df <- data.frame(PassengerID = rep(892:1309),Survived = rf.4.pred)
write.csv(submit.df,file = "/Users/sabbirhassan/Dropbox/ML_stuff/titanic/RF_SUB_20190207_04.csv", row.names = FALSE)
#seems like the less features inceased the test scores
#Lets try crossvalidation 3 fold and try
set.seed(50)
cv.3.folds <- createMultiFolds(rf.label, k = 3, times = 10) #creates bunch of indexes
cntrl.3 <- trainControl(method = "repeatedcv", number = 10, repeats = 10,
index = cv.3.folds)
# setting up doSNOW package for multi-core training; doMC only works on Linux
cl <- makeCluster(4, type = "SOCK")
registerDoSNOW(cl)
set.seed(55)
rf.5.cv.3 <- train(x = rf.train.4, y = rf.label, method = "rf", tuneLength = 3,
ntree = 64, trControl = cntrl.3) #also reducing the trees
#Shut down cluster
stopCluster(cl)
#checkout results
rf.5.cv.3
#so this 3-fold maybe more generalized and better.
test.submit.df <- data.combined[892:1309, features]
rf.5.cv.3.pred <- predict(rf.5.cv.3, test.submit.df)
submit.df <- data.frame(PassengerID = rep(892:1309),Survived = rf.5.cv.3.pred)
write.csv(submit.df,file = "/Users/sabbirhassan/Dropbox/ML_stuff/titanic/RF_SUB_20190207_05.csv", row.names = FALSE)
#some how CV 3 fold doesnt work well. rather just RF is working well
#lets see by CV 10
set.seed(60)
cv.10.folds <- createMultiFolds(rf.label, k = 10, times = 10) #creates bunch of indexes
#check stratification , caret can do it for you.
#its usefull for small amount of imbalancing
cntrl.10 <- trainControl(method = "repeatedcv", number = 10, repeats = 10,
index = cv.10.folds)
# setting up doSNOW package for multi-core training; doMC only works on Linux
cl <- makeCluster(4, type = "SOCK")
registerDoSNOW(cl)
set.seed(65)
rf.5.cv.10 <- train(x = rf.train.4, y = rf.label, method = "rf", tuneLength = 3,
ntree = 64, trControl = cntrl.10) #also reducing the trees
#Shut down cluster
stopCluster(cl)
#checkout results
rf.5.cv.10
test.submit.df <- data.combined[892:1309, features]
rf.5.cv.10.pred <- predict(rf.5.cv.10, test.submit.df)
submit.df <- data.frame(PassengerID = rep(892:1309),Survived = rf.5.cv.10.pred)
write.csv(submit.df,file = "/Users/sabbirhassan/Dropbox/ML_stuff/titanic/RF_SUB_20190207_06.csv", row.names = FALSE)
#using random forest package instead of package
features <- c("Pclass", "New.title", "ticket.party.size", "avg.fare")
rf.train.temp <- data.combined[1:891, features]
set.seed(1234)
rf.temp <- randomForest(x = rf.train.temp, y = rf.label, ntree = 1000)
rf.temp
test.submit.df <- data.combined[892:1309, features]
# Make predictions
rf.preds <- predict(rf.temp, test.submit.df)
table(rf.preds)
# Write out a CSV file for submission to Kaggle
submit.df <- data.frame(PassengerId = rep(892:1309), Survived = rf.preds)
write.csv(submit.df, file = "/Users/sabbirhassan/Dropbox/ML_stuff/titanic/RF_SUB_20190207_07.csv", row.names = FALSE)
# feature analysis very important
# First, let's explore our collection of features using mutual information to
# gain some additional insight. Our intuition is that the plot of our tree
# should align well to the definition of mutual information.
#install.packages("infotheo")
#using mutual information / entropy for feature selection : use stanford link
library(infotheo)
mutinformation(rf.label, data.combined$Pclass[1:891])
mutinformation(rf.label, data.combined$Sex[1:891])
mutinformation(rf.label, data.combined$SibSp[1:891])
mutinformation(rf.label, data.combined$Parch[1:891])
mutinformation(rf.label, discretize(data.combined$Fare[1:891]))
mutinformation(rf.label, data.combined$Embarked[1:891])
mutinformation(rf.label, data.combined$title[1:891])
mutinformation(rf.label, data.combined$family[1:891])
mutinformation(rf.label, data.combined$Ticket.first.char[1:891])
#mutinformation(rf.label, data.combined$Age[1:891])
mutinformation(rf.label, data.combined$New.title[1:891])
mutinformation(rf.label, data.combined$ticket.party.size[1:891])
mutinformation(rf.label, discretize(data.combined$avg.fare[1:891]))
# OK, now let's leverage the tsne algorithm to create a 2-D representation of our data
# suitable for visualization starting with folks our model gets right very often - folks
# with titles other than 'Mr."
#install.packages("Rtsne")
# Also keep dimension reduction in mind. not really needed in trees, but other algorithms might need it.
# PCA, ICA can generate the correspoonding reduced important features.
library #dimensionality reduction library
most.correct <- data.combined[data.combined$New.title != "Mr.",]
indexes <- which(most.correct$Survived != "None")
# NOTE - Bug fix for original version. Rtsne needs a seed to ensure consistent
# output between runs.
set.seed(984357)
tsne.1 <- Rtsne(most.correct[, features], check_duplicates = FALSE)
ggplot(NULL, aes(x = tsne.1$Y[indexes, 1], y = tsne.1$Y[indexes, 2],
color = most.correct$Survived[indexes])) +
geom_point() +
labs(color = "Survived") +
ggtitle("tsne 2D Visualization of Features for new.title Other than 'Mr.'")
# To get a baseline, let's use conditional mutual information on the tsne X and
# Y features for females and boys in 1st and 2nd class. The intuition here is that
# the combination of these features should be higher than any individual feature
# we looked at above.
condinformation(most.correct$Survived[indexes], discretize(tsne.1$Y[indexes,]))
# As one more comparison, we can leverage conditional mutual information using
# the top two features used in our tree plot - new.title and pclass
condinformation(rf.label, data.combined[1:891, c("New.title", "Pclass")])
# OK, now let's take a look at adult males since our model has the biggest
# potential upside for improving (i.e., the tree predicts incorrectly for 86
# adult males). Let's visualize with tsne.
misters <- data.combined[data.combined$New.title == "Mr.",]
indexes <- which(misters$Survived != "None")
set.seed(98437)
tsne.2 <- Rtsne(misters[, features], check_duplicates = FALSE)
ggplot(NULL, aes(x = tsne.2$Y[indexes, 1], y = tsne.2$Y[indexes, 2],
color = misters$Survived[indexes])) +
geom_point() +
labs(color = "Survived") +
ggtitle("tsne 2D Visualization of Features for new.title of 'Mr.'")
# very poor for males misters
# Now conditional mutual information for tsne features for adult males
condinformation(misters$Survived[indexes], discretize(tsne.2$Y[indexes,]))
#
# Idea - How about creating tsne featues for all of the training data and
# using them in our model?
set.seed(987)
tsne.3 <- Rtsne(data.combined[, features], check_duplicates = FALSE)
ggplot(NULL, aes(x = tsne.3$Y[1:891, 1], y = tsne.3$Y[1:891, 2],
color = data.combined$Survived[1:891])) +
geom_point() +
labs(color = "Survived") +
ggtitle("tsne 2D Visualization of Features for all Training Data")
# Now conditional mutual information for tsne features for all training
condinformation(data.combined$Survived[1:891], discretize(tsne.3$Y[1:891,]))
# Add the tsne features to our data frame for use in model building
data.combined$tsne.x <- tsne.3$Y[,1]
data.combined$tsne.y <- tsne.3$Y[,2]
# ----------------------------- #
# we did visualization in training only but we use tsne for both train and test
features <- c("Pclass", "New.title","family","ticket.party.size", "avg.fare","tsne.x","tsne.y")
rf.train.7 <- data.combined[1:891, features]
rf.label <- as.factor(train$Survived)
set.seed(450) #seting seed to verify a random run
rf.7 <- randomForest(x=rf.train.7, y=rf.label, importance = TRUE, ntree = 1000)
rf.7 #just this output gives some info esp the confusion matrix and oob (out of the bag)
importance(rf.7) #getting importance values
varImpPlot(rf.7)
test.submit.df <- data.combined[892:1309, features]
rf.7.pred <- predict(rf.7, test.submit.df)
submit.df <- data.frame(PassengerID = rep(892:1309),Survived = rf.7.pred)
write.csv(submit.df,file = "/Users/sabbirhassan/Dropbox/ML_stuff/titanic/RF_SUB_20190207_08.csv", row.names = FALSE)
|
af508b56e662505139cb4fb1e644d181fc2827d6
|
ffdea92d4315e4363dd4ae673a1a6adf82a761b5
|
/data/genthat_extracted_code/pvsR/examples/CandidateBio.getAddlBio.Rd.R
|
07197f4146c46d8fc42ce7babb9d1f7db4003692
|
[] |
no_license
|
surayaaramli/typeRrh
|
d257ac8905c49123f4ccd4e377ee3dfc84d1636c
|
66e6996f31961bc8b9aafe1a6a6098327b66bf71
|
refs/heads/master
| 2023-05-05T04:05:31.617869
| 2019-04-25T22:10:06
| 2019-04-25T22:10:06
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 569
|
r
|
CandidateBio.getAddlBio.Rd.R
|
library(pvsR)
### Name: CandidateBio.getAddlBio
### Title: Get a candidate's additional biographical information
### Aliases: CandidateBio.getAddlBio
### ** Examples
# First, make sure your personal PVS API key is saved as character string in the pvs.key variable:
## Not run: pvs.key <- "yourkey"
# get additional biographical data on Barack Obama
## Not run: obama <- CandidateBio.getAddlBio(9490)
## Not run: obama
# get additional biographical data on Barack Obama and Mitt Romney
## Not run: onr <- CandidateBio.getAddlBio(list(9490,21942))
## Not run: onr
|
96f35caf5fe5aef409fde95604fc8d7aa484b630
|
bb9140e05d2b493422d65084bc9df4fb6ae88ba9
|
/R/R_cookbook/data_structures/factor_example.R
|
d5ef13aeccd8ecbe771712427ffbb0d25e3d0e37
|
[] |
no_license
|
8589/codes
|
080e40d6ac6e9043e53ea3ce1f6ce7dc86bb767f
|
fd879e36b6d10e5688cc855cd631bd82cbdf6cac
|
refs/heads/master
| 2022-01-07T02:31:11.599448
| 2018-11-05T23:12:41
| 2018-11-05T23:12:41
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 68
|
r
|
factor_example.R
|
f <- factor(c("Win","Win","Lose","Tie","Win","Lose"))
print(f)
wday
|
fbecfa8cdb8c24b2c4d79dd2c2439d6f93ffa635
|
0a906cf8b1b7da2aea87de958e3662870df49727
|
/distr6/inst/testfiles/C_EmpiricalMVPdf/libFuzzer_C_EmpiricalMVPdf/C_EmpiricalMVPdf_valgrind_files/1610036757-test.R
|
9dfed72e18972c277e4bfadc27f585d7957ce119
|
[] |
no_license
|
akhikolla/updated-only-Issues
|
a85c887f0e1aae8a8dc358717d55b21678d04660
|
7d74489dfc7ddfec3955ae7891f15e920cad2e0c
|
refs/heads/master
| 2023-04-13T08:22:15.699449
| 2021-04-21T16:25:35
| 2021-04-21T16:25:35
| 360,232,775
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 330
|
r
|
1610036757-test.R
|
testlist <- list(data = structure(c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0), .Dim = c(10L, 1L)), x = structure(c(1.82391755146545e-183, 1.82391755146545e-183, 8.0988077346472e-179, 5.4674514851239e-304, 4.88059051526537e-312, 0, 8.48798316386109e-314), .Dim = c(7L, 1L)))
result <- do.call(distr6:::C_EmpiricalMVPdf,testlist)
str(result)
|
5b7ba371257b5b0f2a2965eb0b04362cbf50a895
|
32a77ca7d4f4acbc71e8d4c1cdb5fadb9e2a0eef
|
/zarzar_hw06.R
|
3add4fc355bdc8f4bfed9c4b886264d564040d89
|
[] |
no_license
|
ChrisZarzar/quantitative_methods
|
dd2673d839b03206f2ce978f6b1d1c16a16713f0
|
17a4e867c38e32646adbd767b912a5a4be33f848
|
refs/heads/master
| 2020-04-05T09:35:28.117681
| 2018-11-08T20:36:52
| 2018-11-08T20:36:52
| 156,753,903
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 2,112
|
r
|
zarzar_hw06.R
|
#a)
#I am going to plot the PDF
quantiles <- seq(0,5, by =0.01)
pdf.exp <- dexp(quantiles, rate=(1/2))
plot(quantiles, pdf.exp, type='l', main = "PDF: Beta = 2")
#I am going to plot the CDF
quantiles <- seq(0,5, by =0.01)
cdf.exp <- pexp(quantiles, rate=(1/2))
plot(quantiles, cdf.exp, type ='l', main = "CDF: Beta = 2")
#b)
# I need to generate a random exponential dataset with 1500 values at a rate of 3.5
set.seed(35)
exp.dist <- rexp(1500, rate=(1/3.5))
exp.dist
hist(exp.dist)
#c)
#Now I need to sort the data using the sort() function.
sort.exp.dist <- sort(exp.dist)
#Calculate the probability between the 1st and 750th datapoints
(pexp(sort.exp.dist[750],rate=(1/3.5))) - (pexp(sort.exp.dist[1],rate=(1/3.5))) #probability between the 1st to 750th column
#OUTPUT: 0.4768198
#Calculate the probability 50th and 350th datapoints
(pexp(sort.exp.dist[350],rate=(1/3.5))) - (pexp(sort.exp.dist[50],rate=(1/3.5))) ##probability betwee 50th and 350th. You have to subtract the probability from the 1st to the 50th to geth the probability for just tha region between
#OUTPUT: 0.1828813
#Calculate the probability 1000th and 1250th datapoints
(pexp(sort.exp.dist[1250],rate=(1/3.5))) - (pexp(sort.exp.dist[1000],rate=(1/3.5))) #probability betwee 1000th and 1250th. You have to subtract the probability from the 1st to the 1000th to geth the probability for just tha region between
#OUTPUT: 0.183537
#Determine the probability of getting a random value greater than the maximum in the dataset
1-(pexp(sort.exp.dist[1500],rate=(1/3.5)))
#OUTPUT: 0.0005217617
#d)
#Determining the quantile values for the 0.025 and 0.975 probabilities.
#Determine, using ifelse stamtements, what percentage of your random number data fall outside of this range
#What percent should it be?
low.limit <- qexp(0.025,rate=(1/3.5))
low.limit
#OUTPUT: 0.08861233
upper.limit <- qexp(0.975, rate=(1/3.5))
upper.limit
#OUTPUT: 12.91108
low.out.prob <- sum(ifelse(sort.exp.dist<low.limit,1,0))/1500
up.out.prob <- sum(ifelse(sort.exp.dist>upper.limit,1,0))/1500
low.out.prob
#OUTPUT: 0.02533333
up.out.prob
#OUTPUT:0.02933333
|
3cd36d7534bbe2b2de1d4e7e3c05e3c1fc0f586c
|
9b5483c96399f5accf4ee2f8758899f7b41cb5cf
|
/man/clusterability.Rd
|
e6c650723f4668b8ff2cfb3f78cfc070dd061c04
|
[] |
no_license
|
cran/clusterability
|
a7f070fa7b8ddcdd0a9e67b230435c7b3a3ef7d8
|
57a774c8c63c7920b0af55a62db0ea5daa707c85
|
refs/heads/master
| 2020-12-21T21:46:43.838786
| 2020-03-04T10:40:07
| 2020-03-04T10:40:07
| 236,572,725
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| true
| 4,247
|
rd
|
clusterability.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/clusterability.R
\docType{package}
\name{clusterability}
\alias{clusterability}
\alias{clusterability-package}
\title{clusterability: a package to perform tests of clusterability}
\description{
The \code{\link{clusterabilitytest}} function can test for
clusterability of a dataset, and the \code{\link[=print.clusterability]{print}} function
to display output in the console. Below we include code to use with the provided example
datasets. Please see the \code{clusterabilitytest} function for documentation on
available parameters.
}
\examples{
\donttest{
# Normals1
data(normals1)
normals1 <- normals1[,-3]
norm1_dippca <- clusterabilitytest(normals1, "dip")
norm1_dipdist <- clusterabilitytest(normals1, "dip", distance_standardize = "NONE",
reduction = "distance")
norm1_silvpca <- clusterabilitytest(normals1, "silverman", s_setseed = 123)
norm1_silvdist <- clusterabilitytest(normals1, "silverman", distance_standardize = "NONE",
reduction = "distance", s_setseed = 123)
print(norm1_dippca)
print(norm1_dipdist)
print(norm1_silvpca)
print(norm1_silvdist)
# Normals2
data(normals2)
normals2 <- normals2[,-3]
norm2_dippca <-
clusterabilitytest(normals2, "dip")
norm2_dipdist <-
clusterabilitytest(normals2, "dip", reduction = "distance", distance_standardize = "NONE")
norm2_silvpca <- clusterabilitytest(normals2, "silverman", s_setseed = 123)
norm2_silvdist <- clusterabilitytest(normals2, "silverman", reduction = "distance",
distance_standardize = "NONE", s_setseed = 123)
print(norm2_dippca)
print(norm2_dipdist)
print(norm2_silvpca)
print(norm2_silvdist)
# Normals3
data(normals3)
normals3 <- normals3[,-3]
norm3_dippca <- clusterabilitytest(normals3, "dip")
norm3_dipdist <- clusterabilitytest(normals3, "dip", reduction = "distance",
distance_standardize = "NONE")
norm3_silvpca <- clusterabilitytest(normals3, "silverman", s_setseed = 123)
norm3_silvdist <- clusterabilitytest(normals3, "silverman", reduction = "distance",
distance_standardize = "NONE", s_setseed = 123)
print(norm3_dippca)
print(norm3_dipdist)
print(norm3_silvpca)
print(norm3_silvdist)
# Normals4
data(normals4)
normals4 <- normals4[,-4]
norm4_dippca <- clusterabilitytest(normals4, "dip")
norm4_dipdist <- clusterabilitytest(normals4, "dip", reduction = "distance",
distance_standardize = "NONE")
norm4_silvpca <- clusterabilitytest(normals4, "silverman", s_setseed = 123)
norm4_silvdist <- clusterabilitytest(normals4, "silverman", reduction = "distance",
distance_standardize = "NONE", s_setseed = 123)
print(norm4_dippca)
print(norm4_dipdist)
print(norm4_silvpca)
print(norm4_silvdist)
# Normals5
data(normals5)
normals5 <- normals5[,-4]
norm5_dippca <- clusterabilitytest(normals5, "dip")
norm5_dipdist <- clusterabilitytest(normals5, "dip", reduction = "distance",
distance_standardize = "NONE")
norm5_silvpca <- clusterabilitytest(normals5, "silverman", s_setseed = 123)
norm5_silvdist <- clusterabilitytest(normals5, "silverman", reduction = "distance",
distance_standardize = "NONE", s_setseed = 123)
print(norm5_dippca)
print(norm5_dipdist)
print(norm5_silvpca)
print(norm5_silvdist)
# iris
data(iris)
newiris <- iris[,c(1:4)]
iris_dippca <- clusterabilitytest(newiris, "dip")
iris_dipdist <- clusterabilitytest(newiris, "dip", reduction = "distance",
distance_standardize = "NONE")
iris_silvpca <- clusterabilitytest(newiris, "silverman", s_setseed = 123)
iris_silvdist <- clusterabilitytest(newiris, "silverman", reduction = "distance",
distance_standardize = "NONE", s_setseed = 123)
print(iris_dippca)
print(iris_dipdist)
print(iris_silvpca)
print(iris_silvdist)}
# cars
data(cars)
cars_dippca <- clusterabilitytest(cars, "dip")
cars_dipdist <- clusterabilitytest(cars, "dip", reduction = "distance",
distance_standardize = "NONE")
cars_silvpca <- clusterabilitytest(cars, "silverman", s_setseed = 123)
cars_silvdist <- clusterabilitytest(cars, "silverman", reduction = "distance",
distance_standardize = "NONE", s_setseed = 123)
print(cars_dippca)
print(cars_dipdist)
print(cars_silvpca)
print(cars_silvdist)
}
|
aad51b1e6f1515aa0c3af3013e39afaa3acd0035
|
8e8fe47449384105bd58ef569aa509a834494501
|
/Goodies/man/PHS.bxp.Rd
|
4dbb5628637719c0d7892326dc0661fcaffc85f7
|
[] |
no_license
|
Kuvelkar/TEST
|
761d20692383052559081f67abd50293512e875c
|
d06f8c49dfcff128f98ca79f6477c1b1c15a6b11
|
refs/heads/master
| 2021-01-10T16:48:09.896028
| 2016-02-03T08:53:29
| 2016-02-03T08:53:29
| 50,986,445
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 3,564
|
rd
|
PHS.bxp.Rd
|
\name{PHS.bxp}
\alias{PHS.bxp}
\title{Draw Box Plots from Summaries}
\description{Draw box plots based on the given summaries in z. It is usually called from within bowplot.}
\usage{PHS.bxp(z, notch = FALSE, width = NULL, varwidth = FALSE, outline = TRUE,
notch.frac = 0.5, log = "", border = par("fg"), pars = NULL,
frame.plot = axes, horizontal = FALSE, add = FALSE, at = NULL,
show.names = NULL, medlwd = 5, confint = FALSE, confcol = 2,
boxwex = 0.5, staplewex = 1, ...)}
\arguments{
\item{z}{a list containing data summaries to be used in constructing the plots. These are usually the result of a call to boxplot, but can be generated in any fashion.}
\item{notch}{logical if notch is TRUE, a notch is drawn in each side of the boxes. If the notches of two plots do not overlap then the medians are significantly different at the 5 percent level.}
\item{width}{a vector giving the relative widths of the boxes making up the plot.}
\item{varwidth}{logical if varwidth is TRUE, the boxes are drawn with widths proportional to the square-roots of the number of observations in the groups.}
\item{outline}{logical if outline is not true, the outliers are not drawn.}
\item{notch.frac}{numeric in (0,1). When notch = TRUE, the fraction of the box width that the notches should use.}
\item{log}{character, indicating if any axis should be drawn in logarithmic scale.}
\item{border}{character or numeric (vector), the color of the box borders. Is recycled for multiple boxes. Is used as default for the boxcol, medcol, whiskcol, staplecol, and outcol options (see below).}
\item{pars}{graphical parameters.}
\item{frame.plot}{logical, indicating if a ?frame? (box) should be drawn; defaults to TRUE, unless axes = FALSE is specified.}
\item{horizontal}{logical indicating if the boxplots should be horizontal; default FALSE means vertical boxes.}
\item{add}{logical, if true add boxplot to current plot.}
\item{at}{numeric vector giving the locations where the boxplots should be drawn, particularly when add = TRUE; defaults to 1:n where n is the number of boxes.}
\item{show.names}{set to TRUE or FALSE to override the defaults on whether an x-axis label is printed for each group.}
\item{medlwd}{median line width. Setting this parameter implicitly sets the medline parameter to TRUE. The special value, NA, is used to indicate the current line width ( par("lwd")). The default is 5, but the "old" and "att" styles set the it to 5. }
\item{confint}{confidence interval logical flag. If TRUE, use z$conf to display confidence intervals. How the confidence intervals are shown is determined by the confnotch, confcol, confangle and confdensity parameters.}
\item{confcol}{confidence interval color. If supplied, confidence intervals will be filled with the indicated color. The default is 2, but the "old" and "att" styles set it to -1 (no filling). }
\item{boxwex}{box width expansion. The width of the boxes, along with the width of the staples (whisker end caps) and outliers (if drawn as lines), are proportional to this parameter. The default is 0.5, but the "att" and "old" styles set this to 1.}
\item{staplewex}{staple width expansion. Proportional to the box width. The default is 1, but the "old" style sets the default to 0.125.}
\item{\dots}{other arguments.}
}
\details{}
\value{An invisible vector, actually identical to the at argument, with the coordinates ("x" if horizontal is false, "y" otherwise) of box centers, useful for adding to the plot.}
\author{IAZI}
\examples{}
|
365192ce355be889b2dcd0dfada7c68ea3c785fb
|
187337b1f53c771c1537f2dc3b5c6dde99519b02
|
/man/eclat.Rd
|
a58da3317845800ae850374aacfd3619b497725f
|
[] |
no_license
|
lgallindo/arules
|
8da9dabe54a7351afa3e7bec12c0c11bbc9de741
|
887184cd80068e04531b0c13bc231ecbd1afddfb
|
refs/heads/master
| 2023-07-24T02:56:14.691724
| 2018-01-10T19:08:50
| 2018-01-10T19:08:50
| 117,311,210
| 0
| 0
| null | 2018-02-17T08:13:46
| 2018-01-13T03:38:40
|
C
|
UTF-8
|
R
| false
| false
| 2,612
|
rd
|
eclat.Rd
|
\name{eclat}
\alias{eclat}
\title{Mining Associations with Eclat}
\description{
Mine frequent itemsets with the Eclat algorithm.
This algorithm uses simple intersection operations for equivalence
class clustering along with bottom-up lattice traversal.
}
\usage{
eclat(data, parameter = NULL, control = NULL)
}
\arguments{
\item{data}{object of class
\code{\linkS4class{transactions}} or any data structure
which can be coerced into
\code{\linkS4class{transactions}} (e.g., binary
\code{matrix}, \code{data.frame}).}
\item{parameter}{object of class
\code{\linkS4class{ECparameter}} or named list (default
values are: support 0.1 and maxlen 5)}
\item{control}{object of class
\code{\linkS4class{ECcontrol}} or named list for
algorithmic controls.}
}
\details{
Calls the C implementation of the Eclat algorithm by Christian
Borgelt for mining frequent itemsets.
Note for control parameter \code{tidLists=TRUE}:
Since storing transaction ID lists is very memory intensive,
creating transaction ID lists only works for minimum
support values which create a relatively small number of itemsets.
See also \code{\link{supportingTransactions}}.
\code{\link{ruleInduction}} can be used to generate rules from the found itemsets.
A weighted version of ECLAT is available as function \code{\link{weclat}}.
This version can be used to perform weighted association rule mining (WARM).
}
\value{
Returns an object of class \code{\linkS4class{itemsets}}.
}
\references{
Mohammed J. Zaki, Srinivasan Parthasarathy, Mitsunori Ogihara, and Wei
Li. (1997)
\emph{New algorithms for fast discovery of association rules}.
Technical Report 651, Computer Science Department, University of
Rochester, Rochester, NY 14627.
Christian Borgelt (2003) Efficient Implementations of Apriori and
Eclat. \emph{Workshop of Frequent Item Set Mining Implementations}
(FIMI 2003, Melbourne, FL, USA).
ECLAT Implementation: \url{http://www.borgelt.net/eclat.html}
}
\seealso{
\code{\link{ECparameter-class}},
\code{\link{ECcontrol-class}},
\code{\link{transactions-class}},
\code{\link{itemsets-class}},
\code{\link{weclat}},
\code{\link{apriori}},
\code{\link{ruleInduction}},
\code{\link{supportingTransactions}}
}
\author{Michael Hahsler and Bettina Gruen}
\examples{
data("Adult")
## Mine itemsets with minimum support of 0.1 and 5 or less items
itemsets <- eclat(Adult,
parameter = list(supp = 0.1, maxlen = 5))
itemsets
## Create rules from the itemsets
rules <- ruleInduction(itemsets, Adult, confidence = .9)
rules
}
\keyword{models}
|
f7d4c936c395e5cc69130c02bde5f88053d3cb9c
|
570f57c6d5355d2064d123e4452094090f16d5b9
|
/Assignment 2 code.R
|
08b00730becf8df9ff892a2e766ec5dda8cf5a4a
|
[] |
no_license
|
R-pidit/R-projects
|
7227aa6d5f45d841e53289f7a41eb9daed1491d2
|
6ac3f271fed69ce54b135b4b554a37085099ffee
|
refs/heads/master
| 2021-01-22T14:55:51.562359
| 2017-09-07T11:09:20
| 2017-09-07T11:09:20
| 102,372,084
| 5
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 4,475
|
r
|
Assignment 2 code.R
|
install.packages("titanic")
install.packages("rpart.plot")
install.packages("randomForest")
install.packages("DAAG")
library(titanic)
library(rpart.plot)
library(gmodels)
library(Hmisc)
library(pROC)
library(ResourceSelection)
library(car)
library(caret)
library(dplyr)
library(InformationValue)
library(rpart)
library(randomForest)
library("DAAG")
cat("\014") # Clearing the screen
getwd()
setwd("C:\\Users\\Subha\\Documents\\R_Data") #This working directory is the folder where all the bank data is stored
rm(list = ls())
titanic_data<-read.csv('train.csv')
titanic_train= titanic_data[c("Pclass" ,"Sex" ,"Age" ,"SibSp", "Parch","Survived")]
#titanic test
titanic_test <-read.csv('test-3.csv')
# number of survived vs number of dead
CrossTable(titanic_train$Survived)
View(titanic_train)
# replacing NA in age column by it's mean
titanic_train$Age[is.na(titanic_train$Age)]= mean(titanic_train$Age[!is.na(titanic_train$Age)])
summary(titanic_train)
#splitting titanic train into 70,30
set.seed(1234) # for reproducibility
titanic_train$rand <- runif(nrow(titanic_train))
titanic_train_start <- titanic_train[titanic_train$rand <= 0.7,]
titanic_test_start <- titanic_train[titanic_train$rand > 0.7,]
View(titanic_train_start)
########## Model building ##########
full.model.titanic.mean <- glm(formula = Survived ~ Pclass+Sex+Age+SibSp+Parch,data = titanic_train_start,family = binomial) #family = binomial implies that it is logistic regression
summary(full.model.titanic.mean)
#removing insignificant variables
titanic_train_start$Parch<-NULL
full.model.titanic.mean <- glm(formula = Survived ~ Pclass+Sex+Age+SibSp,
data=titanic_train_start, family = binomial) #family = binomial implies that the type of regression is logistic
summary(full.model.titanic.mean)
#All variables significant
#Testing performance on Train set
titanic_train_start$prob = predict(full.model.titanic.mean, type=c("response"))
titanic_train_start$Survived.pred = ifelse(titanic_train_start$prob>=.5,'pred_yes','pred_no')
table(titanic_train_start$Survived.pred,titanic_train_start$Survived)
#Testing performance on test set
nrow(titanic_test)
titanic_test_start$prob = predict(full.model.titanic.mean, newdata=titanic_test_start, type=c("response"))
titanic_test_start$Survived.pred = ifelse(titanic_test_start$prob>=.5,'pred_yes','pred_no')
table(titanic_test_start$Survived.pred,titanic_test_start$Survived)
########## END - Model with mean included instead of NA #########
### Testing for Jack n Rose's survival ###
df.jackrose <- read.csv('Book1.csv')
df.jackrose$prob = predict(full.model.titanic.mean, newdata=df.jackrose, type=c("response"))
df.jackrose$Survived.pred = ifelse(df.jackrose$prob>=.5,'pred_yes','pred_no')
head(df.jackrose)
# Jack dies, Rose survives
### END - Testing on Jack n Rose ###
## START K-fold cross validation ##
# Defining the K Fold CV function here
Kfold_func <- function(dataset,formula,family,k)
{
object <- glm(formula=formula, data=dataset, family = family)
CVbinary(object, nfolds= k, print.details=TRUE)
}
#Defining the function to calculate Mean Squared Error here
MeanSquareError_func <- function(dataset,formula)
{
LM_Object <- lm(formula=formula, data=dataset)
LM_Object_sum <-summary(LM_Object)
MSE <- mean(LM_Object_sum$residuals^2)
print("Mean squared error")
print(MSE)
}
#Performing KFold CV on Training set by calling the KFOLD CV function here
Kfoldobj <- Kfold_func(titanic_train_start,Survived ~ Pclass + Sex + SibSp + Age,binomial,10)
#Calling the Mean Squared Error function on the training set here
MSE_Train <-MeanSquareError_func(titanic_train_start,Survived ~ Pclass + Sex + SibSp + Age)
#confusion matrix on training set
table(titanic_train_start$Survived,round(Kfoldobj$cvhat))
print("Estimate of Accuracy")
print(Kfoldobj$acc.cv)
#Performing KFold CV on test set by calling the KFOLD CV function here
Kfoldobj.test <- Kfold_func(titanic_test_start,Survived ~ Pclass + Sex + SibSp + Age,binomial,10)
#Calling the Mean Squared Error function on the test set here
MSE_Test <-MeanSquareError_func(titanic_test_start,Survived ~ Pclass + Sex + SibSp + Age)
#Confusion matrix on test set
table(titanic_test_start$Survived,round(Kfoldobj.test$cvhat))
print("Estimate of Accuracy")
print(Kfoldobj.test$acc.cv)
## END K-FOLD CROSS VALIDATION ##
|
ece440a44978c2ff0932d186df772f1bd402e513
|
8bde36c00a458f1d3b3c81eea824c56bc72b26e2
|
/approximate_UMVUE.R
|
cf57a4d4fd65278e0e72e433d3927c47406a6cea
|
[] |
no_license
|
snigdhagit/Bayesian-selective-inference
|
ea186c866700510cf5206977e936e6224dbc5dd2
|
a4e436aef0636c6acc6385e75f40109ead722d21
|
refs/heads/master
| 2021-07-08T07:45:00.242158
| 2017-10-02T08:47:00
| 2017-10-02T08:47:00
| 105,512,054
| 1
| 2
| null | null | null | null |
UTF-8
|
R
| false
| false
| 544
|
r
|
approximate_UMVUE.R
|
#computes approximate UMVUE under additive Gaussian randomization with variance tau^2 (in the univariate case)
UMVUE.compute<-function(y, sigma, tau)
{
objective<-function(z,alpha)
{
(((-alpha*z))+((z^2/2)+log(1+(1/z))))
}
objective1<-function(z,alpha)
{
(((-alpha*z)/tau)+((z^2/2)+log(1+(1/z))))
}
approx1<-function(alpha1)
{
opt<-nlm(objective1,p=1,alpha=alpha1,hessian=TRUE)
return(opt$estimate)
}
UMVUE.approx<-(y*(1+((sigma^2)/(tau^2))))-(((sigma^2)/(tau))*approx1(y))
return(UMVUE.approx)
}
|
f3a5851977bbf960cc10042ff7fcd3375fd35624
|
bccaf9ca75d67fef6bec733e784c582149a32ed1
|
/plagiat/R/jclu2bup.f.R
|
89e1438d0b95b2f91986de64fd67d7cdb98df4c2
|
[] |
no_license
|
brooksambrose/pack-dev
|
9cd89c134bcc80711d67db33c789d916ebcafac2
|
af1308111a36753bff9dc00aa3739ac88094f967
|
refs/heads/master
| 2023-05-10T17:22:37.820713
| 2023-05-01T18:42:08
| 2023-05-01T18:42:08
| 43,087,209
| 0
| 1
| null | null | null | null |
UTF-8
|
R
| false
| false
| 1,050
|
r
|
jclu2bup.f.R
|
#' JSTOR Clustering 2 Bimodal Unimodal Projection Illustration
#'
#' @param jclu
#' @param s
#'
#' @return
#' @export
#' @import igraph magrittr data.table
#'
#' @examples
jclu2bup.f<-function(jclu,s=300){
par(mar=c(0,0,1,0))
bs<-igraph::as_edgelist(jclu$b) %>% data.table %>% setnames(ec('j,l')) %>% .[sample(1:.N,s)] %>% as.matrix %>% graph_from_edgelist
V(bs)$type<-grepl('^[A-Z]',V(bs)$name)
bs$name<-'journal-label'
args<-function(x) list(x=x,main=graph_attr(x, 'name'),label.family='serif',vertex.label=NA,vertex.size=4,edge.arrow.size=0,vertex.color=V(x)$type+1,vertex.frame.color=NA,edge.color=if(E(x)$weight %>% is.null) gray(.2) else sapply(E(x)$weight,function(y) gray(level=1-(y/max(E(x)$weight)))))
ms<-bipartite_projection(bs,remove.type = F)
ms$proj1$name<-'journal'
ms$proj2$name<-'label'
docx<-'docx'%in%knitr::opts_knit$get("rmarkdown.pandoc.to")
if(docx) par(mfrow=c(1,3))
do.call(plot,args(bs))
do.call(plot,args(ms$proj1))
do.call(plot,args(ms$proj2))
if(docx) par(mfrow=c(1,1))
c(list(b=bs),ms)
}
|
7a8a8347270b7c5bfd44f40a1de964a87bab6a5f
|
bb767f6a07340c0c313c79587ea6c96ce5e17f33
|
/R/data.r
|
da1aa212d98a75fe535d3ca11d467ba2b7815e10
|
[] |
no_license
|
psychobas/corpustools
|
82694086aa0b3d861e38624a5cf17a53ce61e23e
|
e9c1ac2011234a62b2fc2c7b46ab01dd9159e4ac
|
refs/heads/master
| 2023-05-04T01:35:30.735603
| 2021-05-25T10:41:30
| 2021-05-25T10:41:30
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 808
|
r
|
data.r
|
#' State of the Union addresses
#'
#' @docType data
#' @usage data(sotu_texts)
#' @format data.frame
'sotu_texts'
##save(sotu_texts, file='data/sotu_texts.rda', compression_level = 9)
#' coreNLP example sentences
#'
#' @docType data
#' @usage data(corenlp_tokens)
#' @format data.frame
'corenlp_tokens'
#' A tCorpus with a small sample of sotu paragraphs parsed with udpipe
#'
#' @docType data
#' @usage data(tc_sotu_udpipe)
#' @format data.frame
'tc_sotu_udpipe'
## run if tc methods have been updated
## tc_sotu_udpipe = refresh_tcorpus(tc_sotu_udpipe)
## save(tc_sotu_udpipe, file='data/tc_sotu_udpipe.rda', compression_level = 9)
#' Basic stopword lists
#'
#' @docType data
#' @usage data(stopwords_list)
#' @format A named list, with names matching the languages used by SnowballC
"stopwords_list"
|
d165b85aafed9b2a16ca2e7d132720bc7eca0e02
|
eff4f65785cdd0f198245b46876720ee4cf40bba
|
/Intro to R II.R
|
fb055ab99c6c1967bde355baa1e7f89b8fa3be7e
|
[] |
no_license
|
PRATIKSHIRBHATE/r_training
|
afc84e0f5a24ded8c4cbb00c8499bd5eb1605e14
|
d11dbe97ac602eaca0c37c1285dcdd4cfe623f47
|
refs/heads/master
| 2020-09-25T15:53:36.079049
| 2019-12-05T11:13:50
| 2019-12-05T11:13:50
| 226,038,433
| 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 7,153
|
r
|
Intro to R II.R
|
## Introduction to R - Part II
## This file contains all of the same code that is contained in the Markdown
## version of the guide. This document serves to demonstrate how a regular R
## script looks and how to run code in this format.
# Loading the packages and data as shown on the previous guide
# "Introduction to R".
packages.to.load <- c("plyr", "dplyr" ,"DT", "ggplot2", "plotly", "RODBC")
invisible(lapply(packages.to.load, library, character.only=TRUE))
load("C:/Users/shregmi/Documents/Current Projects/R Training/Example Datasets/Auto.rda")
write.csv(Auto, file="Autodata.csv")
auto.data <- read.csv(file ="C:/Users/shregmi/Documents/Current Projects/R Training/Autodata.csv"
, header = TRUE)
#Manipulating Data
# This section will cover a couple ways to shape data into something that is
# easy to work with.
##Using the $ operator
#For very simple filters and selection of certain columns from a data frame, you
# can use the $ operator. Here, we take one column from our auto.data data frame
# and assign it to a new object. You can use the class() function to see what
# kind of object it is.
JustMPG <- matrix(auto.data$mpg)
class(JustMPG)
# We can also pull multiple columns pretty easily.
MPGandWeight <- data.frame(auto.data$mpg, auto.data$weight)
head(MPGandWeight)
##dplyr
# The package dplyr is a heavily used library for data cleaning and preparation
# (one of R's main strong suits). Here is a little cheat sheet covering some
# dplyr and tidyr features:
# https://www.rstudio.com/wp-content/uploads/2015/02/data-wrangling-cheatsheet.pdf
# The official dplyr documentation can also be found here:
# https://cran.r-project.org/web/packages/dplyr/dplyr.pdf
#There are 5 main functions or "verbs" used in dplyr: Select, Filter, Mutate,
# Arrange, and Summarize. Many of these align with the main SQL commands:
# Select, Where, Group By, etc...
###Select
# Let's take a subset of columns from the auto.data dataset.
auto.subset <- select(auto.data, name, mpg, cylinders, weight)
head(auto.subset)
# We can also specify which columns to exclude using the "-" operator.
auto.subset2 <- select(auto.data, -horsepower, -year)
head(auto.subset2)
###Filter
# After selecting your data of interest, you can pass filters in the same way as
# you would a "where" clause in SQL.
only.4cyl <- filter(auto.subset, cylinders==4)
head(only.4cyl)
# You can add multiple arguments to the filter function.
multi.filter <- filter(auto.subset, cylinders %in% c(4,6), weight<3000)
head(multi.filter)
# We can filter strings as well.
filter.name <- filter(auto.subset, cylinders==4, weight<2000,
grepl("toyota|volkswagen", name))
head(filter.name)
###Mutate
# Using mutate, we can create new columns. Additionally, we can chain the
# previous "verbs" together using the "%>%" operator (called pipe operator).
# Note that the dplyr package must be loaded in order to use this operator.
add.col <- auto.data %>%
select(name, mpg, cylinders, weight) %>%
filter(weight>1800) %>%
mutate(weightpercyl <- weight/cylinders)
head(add.col)
###Arrange
# This is equivalent to an "order by" statement in SQL. Here, we sort the
# previous output by "weightpercyl" in descending order using the "-" operator.
add.col.sorted <- auto.data %>%
select(name, mpg, cylinders, weight) %>%
filter(weight>1800) %>%
mutate(weightpercyl = weight/cylinders) %>%
arrange(-weightpercyl)
head(add.col.sorted)
###Summarize
summarized.auto <- auto.data %>%
select(name, mpg, cylinders, weight) %>%
filter(weight>1800) %>%
mutate(weightpercyl = weight/cylinders) %>%
summarise(avg_mpg = mean(mpg),
min_weight = min(weight),
median_weightpercyl = median(weightpercyl))
summarized.auto
# We can pass a group_by() clause as well.
summarized.auto2 <- auto.data %>%
select(name, mpg, cylinders, weight) %>%
filter(weight>1800) %>%
mutate(weightpercyl = weight/cylinders) %>%
group_by(cylinders) %>%
summarise(avg_mpg = mean(mpg),
avg_weight = mean(weight),
avg_weightpercyl = mean(weightpercyl))
summarized.auto2
#Visualization
# This section will cover a few useful packages and methods for visualization in
# an R Markdown document.
##Displaying Tables in R Markdown
# When using print() or head() functions to show the contents of a data frame,
# the output looks very ugly. There are a huge number of packages that are
# specifically for displaying tables in an aesthetically pleasing way. The "DT"
# or Data Tables package is one of them.
datatable(auto.data)
# This is the default way to display a data frame with this package with no
# arguments provided except for the object that is to be shown.
# There are a huge number of arguments you can add to make your reports look
# nicer and to add functionality. BONUS: You'll also write the smallest piece of
# JavaScript code in this as well.
datatable(auto.data, extensions = 'Buttons', options = list(
dom = 'Bfrtip',
buttons= c('copy', 'excel'),
pageLength = 5,
initComplete = JS("
function(settings, json) {
$(this.api().table().header()).css({
'background-color': '#232D69',
'color': '#fff'
});
}")
), rownames = FALSE)
##Charts with Plotly
# Cheat sheet for plotly:
# https://images.plot.ly/plotly-documentation/images/r_cheat_sheet.pdf.
# Another great resource: https://plot.ly/r/
plot <- plot_ly(x = auto.data$weight, type="histogram")
plot
# In plotly, you can add multiple sets of traces or bars to the same plot with
# the use of pipe operators. You can also name each of the traces/bars so that
# the result is clear.
plot2 <- plot_ly(data =auto.data, alpha = 0.5) %>% #adjusting color transparency
add_histogram(x = auto.data$weight[auto.data$origin == 1], name="American") %>%
add_histogram(x = auto.data$weight[auto.data$origin == 2], name="European") %>%
add_histogram(x = auto.data$weight[auto.data$origin == 3], name="Japanese") %>%
layout(barmode = "overlay")
plot2
# Here is an example of a scatterplot generated by plotly. This has many more
# features and much more functionality than the plot we first generated with the
# R Base plot() function.
auto.data$OriginName[which(auto.data$origin == 1)] = "American"
auto.data$OriginName[which(auto.data$origin == 2)] = "European"
auto.data$OriginName[which(auto.data$origin == 3)] = "Japanese"
plot3 <- plot_ly(data = auto.data
,x = ~weight
,y = ~mpg
,color = ~OriginName
,type ="scatter"
,mode="markers"
,text= ~paste("Car name: ", name
,"</br> Year: ", year
,"</br> Cylinders: ", cylinders))
plot3
##Charts with ggplot2
#Cheat sheet for ggplot2: https://www.rstudio.com/wp-content/uploads/2015/03/ggplot2-cheatsheet.pdf
|
38ac95555fe279ef8ae6c775d78ed81c1f14e47d
|
ffdea92d4315e4363dd4ae673a1a6adf82a761b5
|
/data/genthat_extracted_code/HelpersMG/examples/modeled.hist.Rd.R
|
271dde27ceb26963b9b570054bf6ba2ea8b265bf
|
[] |
no_license
|
surayaaramli/typeRrh
|
d257ac8905c49123f4ccd4e377ee3dfc84d1636c
|
66e6996f31961bc8b9aafe1a6a6098327b66bf71
|
refs/heads/master
| 2023-05-05T04:05:31.617869
| 2019-04-25T22:10:06
| 2019-04-25T22:10:06
| null | 0
| 0
| null | null | null | null |
UTF-8
|
R
| false
| false
| 615
|
r
|
modeled.hist.Rd.R
|
library(HelpersMG)
### Name: modeled.hist
### Title: Return the theoretical value for the histogram bar
### Aliases: modeled.hist
### ** Examples
## Not run:
##D n <- rnorm(100, mean=10, sd=2)
##D breaks <- 0:20
##D hist(n, breaks=breaks)
##D
##D s <- modeled.hist(breaks=breaks, FUN=pnorm, mean=10, sd=2, sum=100)
##D
##D points(s$x, s$y, pch=19)
##D lines(s$x, s$y)
##D
##D n <- rlnorm(100, meanlog=2, sdlog=0.4)
##D b <- hist(n, ylim=c(0, 70))
##D
##D s <- modeled.hist(breaks=b$breaks, FUN=plnorm, meanlog=2, sdlog=0.4, sum=100)
##D
##D points(s$x, s$y, pch=19)
##D lines(s$x, s$y)
## End(Not run)
|
dcc6d38fe84ec6562e1230189687a50a97624fa0
|
ef424746a3ea4ed6e167f03d359b39da48a0fc21
|
/man/colLuminosity_utility.Rd
|
a6badc261d6881eba0e7c11abd380faa9f24fb26
|
[] |
no_license
|
smitdave/MASH
|
397a1f501c664089ea297b8841f2cea1611797e4
|
b5787a1fe963b7c2005de23a3e52ef981485f84c
|
refs/heads/master
| 2021-01-18T18:08:25.424086
| 2017-08-17T00:18:52
| 2017-08-17T00:18:52
| 86,845,212
| 0
| 3
| null | 2017-08-17T00:18:52
| 2017-03-31T17:42:46
|
R
|
UTF-8
|
R
| false
| true
| 671
|
rd
|
colLuminosity_utility.Rd
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/MICRO-Landscape-Utilities.R
\name{colLuminosity_utility}
\alias{colLuminosity_utility}
\title{Brighten or Darken Colors}
\usage{
colLuminosity_utility(color, factor, bright, alpha = NULL)
}
\arguments{
\item{color}{vector of hcl colors}
\item{factor}{factor to brighten or darken colors}
\item{bright}{logical variable to brighten or darken}
\item{alpha}{opacity}
}
\value{
a vector of colors in hex format
}
\description{
With input of hcl colors (hex code), brighten or darken by a factor
}
\examples{
colLuminosity_utility(color=MASH::ggCol_utility(n=5), factor = 1.15, bright = TRUE)
}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.