Id stringlengths 1 6 | PostTypeId stringclasses 7
values | AcceptedAnswerId stringlengths 1 6 ⌀ | ParentId stringlengths 1 6 ⌀ | Score stringlengths 1 4 | ViewCount stringlengths 1 7 ⌀ | Body stringlengths 0 38.7k | Title stringlengths 15 150 ⌀ | ContentLicense stringclasses 3
values | FavoriteCount stringclasses 3
values | CreationDate stringlengths 23 23 | LastActivityDate stringlengths 23 23 | LastEditDate stringlengths 23 23 ⌀ | LastEditorUserId stringlengths 1 6 ⌀ | OwnerUserId stringlengths 1 6 ⌀ | Tags list |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2711 | 1 | null | null | 4 | 146 | I'm trying to calculate how good are my measurements in machine learning! Let's say that I have five choices, and that error is 4, 2, 0.002, 3, 6. Naturally, I will pick third one for the hit, but I would like to say following:
- I'm X% certain that hit is third pick
- I'm Y% certain that hit is first (last) pick
... | How good is my error? | CC BY-SA 2.5 | null | 2010-09-16T08:01:38.570 | 2010-10-17T11:45:39.267 | 2010-09-17T11:04:45.970 | 159 | 1313 | [
"machine-learning",
"model-selection"
] |
2712 | 2 | null | 2711 | 2 | null | I would recommend cross-validation; then, from all iterations, you can get not only average, but also an approximation of distribution of error, and you can use it to estimate the confidence interval.
| null | CC BY-SA 2.5 | null | 2010-09-16T09:24:21.157 | 2010-09-16T09:24:21.157 | null | null | null | null |
2713 | 2 | null | 2272 | 15 | null | The answers provided before are very helpful and detailed. Here is my $0.25.
Confidence interval (CI) is a concept based on the classical definition of probability (also called the "Frequentist definition") that probability is like proportion and is based on the axiomatic system of Kolmogrov (and others).
Credible int... | null | CC BY-SA 4.0 | null | 2010-09-16T09:35:44.257 | 2018-11-18T23:03:07.730 | 2018-11-18T23:03:07.730 | 79696 | 1307 | null |
2714 | 2 | null | 1487 | 0 | null | Using PCA only to analyze your gene expression data can lead to (possibly) wrong conclusions. The fundamental assumption behind PCA is equal genewise variation and this may NOT be justified almost always. The links to the paper to analyze time course data by csgillespie point to better ways.
| null | CC BY-SA 2.5 | null | 2010-09-16T10:02:46.633 | 2010-09-16T10:02:46.633 | null | null | 1307 | null |
2715 | 1 | null | null | 88 | 10508 | I like G van Belle's book on [Statistical Rules of Thumb](http://vanbelle.org/), and to a lesser extent [Common Errors in Statistics (and How to Avoid Them)](http://eu.wiley.com/WileyCDA/WileyTitle/productCd-1118294394.html) from Phillip I Good and James W. Hardin. They address common pitfalls when interpreting results... | Rules of thumb for "modern" statistics | CC BY-SA 3.0 | null | 2010-09-16T10:21:36.837 | 2019-04-12T22:15:09.463 | 2017-01-17T11:02:13.353 | 96626 | 930 | [
"modeling",
"exploratory-data-analysis",
"rule-of-thumb"
] |
2716 | 2 | null | 2691 | 36 | null | OK, a totally non-math answer:
If you have a bunch of variables on a bunch of subjects and you want to reduce it to a smaller number of variables on those same subjects, while losing as little information as possible, then PCA is one tool to do this.
It differs from factor analysis, although they often give similar res... | null | CC BY-SA 2.5 | null | 2010-09-16T10:23:42.837 | 2010-09-16T10:23:42.837 | null | null | 686 | null |
2717 | 1 | 2721 | null | 72 | 95730 | I have a (symmetric) matrix `M` that represents the distance between each pair of nodes. For example,
A B C D E F G H I J K L
A 0 20 20 20 40 60 60 60 100 120 120 120
B 20 0 20 20 60 80 80 80 120 140 140 140
C 20 20 0 20 60 80 80 80 120 140 140 140
D 20 20 20 0... | Clustering with a distance matrix | CC BY-SA 2.5 | null | 2010-09-16T11:47:15.633 | 2023-03-25T22:33:04.297 | 2023-03-25T22:33:04.297 | 60613 | 1316 | [
"clustering",
"distance-functions"
] |
2718 | 2 | null | 2715 | 38 | null |
## There is no free lunch
A large part of statistical failures is created by clicking a big shiny button called "Calculate significance" without taking into account its burden of hidden assumptions.
## Repeat
Even if a single call to a random generator is involved, one may have luck or bad luck and so jump to the... | null | CC BY-SA 3.0 | null | 2010-09-16T12:08:07.737 | 2013-06-17T18:04:25.600 | 2020-06-11T14:32:37.003 | -1 | null | null |
2719 | 2 | null | 2715 | 18 | null | There can be a long list but to mention a few: (in no specific order)
- P-value is NOT probability. Specifically, it is not the probability of committing Type I error. Similarly, CIs have no probabilistic interpretation for the given data. They are applicable for repeated experiments.
- Problem related to variance do... | null | CC BY-SA 3.0 | null | 2010-09-16T12:15:45.783 | 2017-07-08T08:12:59.197 | 2017-07-08T08:12:59.197 | 79696 | 1307 | null |
2720 | 2 | null | 1815 | 9 | null | I am surprise no one mentioned: Statistical Design by George Casella
[Google Books Link](http://books.google.com/books?id=sqnbSUtryVAC&printsec=frontcover&dq=statistical++design+casella+book&source=bl&ots=634WifKYf5&sig=nfzDyBldgGgNFTc8rl8i8XYLfp8&hl=en&ei=sQ2STNnXG4Oenwf06_nVCA&sa=X&oi=book_result&ct=result&resnum=9&v... | null | CC BY-SA 2.5 | null | 2010-09-16T12:31:45.183 | 2010-09-16T12:31:45.183 | null | null | 1307 | null |
2721 | 2 | null | 2717 | 48 | null | There are a number of options.
## k-medoids clustering
First, you could try partitioning around medoids (pam) instead of using k-means clustering. This one is more robust, and could give better results. Van der Laan reworked the algorithm. If you're going to implement it yourself, his [article](https://biostats.bepr... | null | CC BY-SA 4.0 | null | 2010-09-16T12:36:49.757 | 2019-12-17T11:50:35.440 | 2019-12-17T11:50:35.440 | 142758 | 1124 | null |
2722 | 2 | null | 1815 | 2 | null | [The Design of Experiments: Statistical Principles for Practical Applications](http://books.google.co.uk/books?id=CaFZPbCllrMC) by Roger Mead. Examples are drawn from agriculture and biology, so probably most appropriate if you're interested in one of those fields. Rather expensive for a 600-page paperback but you can ... | null | CC BY-SA 2.5 | null | 2010-09-16T12:45:22.580 | 2010-09-16T12:45:22.580 | null | null | 449 | null |
2723 | 2 | null | 2715 | 66 | null | Don't forget to do some basic data checking before you start the analysis. In particular, look at a scatter plot of every variable you intend to analyse against ID number, date / time of data collection or similar. The eye can often pick up patterns that reveal problems when summary statistics don't show anything unusu... | null | CC BY-SA 2.5 | null | 2010-09-16T12:57:26.753 | 2010-09-16T12:57:26.753 | null | null | 449 | null |
2724 | 2 | null | 2715 | 29 | null | One thing I tell my students is to produce an appropriate graph for every p-value. e.g., a scatterplot if they test correlation, side-by-side boxplots if they do a one-way ANOVA, etc.
| null | CC BY-SA 2.5 | null | 2010-09-16T13:13:10.500 | 2010-09-16T13:13:10.500 | null | null | 159 | null |
2726 | 1 | 2735 | null | 5 | 181 | Is there a way to have R plot to an in-memory object or connection, rather than a named file?
I would like to have a plotting server create many graphs without ever going to a file.
The [Cairo](http://cran.r-project.org/web/packages/Cairo/index.html) package documents use of a connection, but it doesn't seem to work. ... | R plots to a connection? | CC BY-SA 2.5 | null | 2010-09-16T13:46:18.170 | 2010-09-16T16:46:17.977 | null | null | 1119 | [
"r",
"data-visualization"
] |
2727 | 2 | null | 2717 | 26 | null | One way to highlight clusters on your distance matrix is by way of [Multidimensional scaling](http://en.wikipedia.org/wiki/Multidimensional_scaling). When projecting individuals (here what you call your nodes) in an 2D-space, it provides a comparable solution to PCA. This is unsupervised, so you won't be able to specif... | null | CC BY-SA 2.5 | null | 2010-09-16T13:59:10.410 | 2010-09-16T13:59:10.410 | 2017-04-13T12:44:55.630 | -1 | 930 | null |
2728 | 1 | 2738 | null | 6 | 3119 | I am clustering a dataset using the pam command (from {cluster} package), and I wish to decide on the number of clusters to use.
I was able to implement The_Elbow_Method in R ([see wiki](http://en.wikipedia.org/wiki/Determining_the_number_of_clusters_in_a_data_set#The_Elbow_Method)) for doing that. But that doesn't pr... | Algorithm for choosing the number of clusters when using pam in R? | CC BY-SA 2.5 | null | 2010-09-16T15:40:06.907 | 2010-11-18T20:58:25.817 | 2010-09-16T16:20:44.223 | null | 253 | [
"r",
"clustering",
"model-selection"
] |
2729 | 2 | null | 2728 | 3 | null | What about [silhouette](http://bm2.genes.nig.ac.jp/RGM2/R_current/library/cluster/html/silhouette.html)?
| null | CC BY-SA 2.5 | null | 2010-09-16T15:43:10.977 | 2010-09-16T15:43:10.977 | null | null | null | null |
2730 | 1 | 2778 | null | 24 | 9986 | I am running experiments for a paper and I am looking for an interesting book / website to understand properly how ANOVA and ANCOVA work. I have a good math background so I don't necessarily need a vulgarized explanation.
I'd also like to know how to determine when to use ANOVA instead of ANCOVA.
| Good resource to understand ANOVA and ANCOVA? | CC BY-SA 3.0 | null | 2010-09-16T15:53:31.090 | 2022-10-04T13:56:21.240 | 2015-10-07T13:55:52.557 | 84004 | 1320 | [
"anova",
"references",
"ancova"
] |
2731 | 2 | null | 2728 | 4 | null | The [fpc package](http://cran.r-project.org/web/packages/fpc/index.html) provides a few clustering statistics. If you're looking for information criteria in particular, the `cluster.stats` method provides an information based distance. For mixture models based on clustering, the BIC is available.
| null | CC BY-SA 2.5 | null | 2010-09-16T16:02:54.897 | 2010-09-16T16:02:54.897 | null | null | 251 | null |
2732 | 2 | null | 2730 | 19 | null | So, in addition to this paper, [Misunderstanding Analysis of Covariance](https://web.archive.org/web/20120324142111/http://mres.gmu.edu/pmwiki/uploads/Main/ancova.pdf), which enumerates common pitfalls when using ANCOVA, I would recommend starting with:
- Frank Harrell's homepage, especially his handout on Regression ... | null | CC BY-SA 4.0 | null | 2010-09-16T16:06:13.050 | 2022-10-04T13:56:21.240 | 2022-10-04T13:56:21.240 | 7290 | 930 | null |
2733 | 2 | null | 485 | 4 | null | There are a bunch of helpful video tutorials on basic statistics & data mining with R and Weka at SentimentMining.net.
- http://sentimentmining.net/StatisticsVideos/
| null | CC BY-SA 2.5 | null | 2010-09-16T16:27:02.927 | 2010-09-16T16:27:02.927 | null | null | 653 | null |
2734 | 2 | null | 2730 | 4 | null | [The R book](http://rads.stackoverflow.com/amzn/click/0470510242) does a good job on that. You can see that it dedicates one chapter to each one of those methods (11 and 12). If you are new to R, this is a great book to start with.
| null | CC BY-SA 2.5 | null | 2010-09-16T16:37:35.490 | 2010-09-16T16:48:14.703 | 2010-09-16T16:48:14.703 | 339 | 339 | null |
2735 | 2 | null | 2726 | 3 | null | I don't think so; R does not have binary memory buffers. Cairo feature you mentioned needs recompilation of both R and package; trying to do this with plain R does not throw errors, but nothing is written neither to textConnection nor socket.
So I think the best idea will be to use ramdisk; if you still want to try the... | null | CC BY-SA 2.5 | null | 2010-09-16T16:46:17.977 | 2010-09-16T16:46:17.977 | null | null | null | null |
2736 | 2 | null | 1736 | 8 | null | [PyIMSL](http://www.vni.com/campaigns/pyimslstudioeval/) contains a handful of routines for survival analyses. It is Free As In Beer for noncommercial use, fully supported otherwise. From the documentation in the Statistics User Guide...
Computes Kaplan-Meier estimates of survival
probabilties: kaplanMeierEstimates()... | null | CC BY-SA 2.5 | null | 2010-09-16T16:53:28.610 | 2010-09-16T18:58:26.863 | 2010-09-16T18:58:26.863 | 1080 | 1080 | null |
2737 | 2 | null | 2686 | 1 | null | I have heard of 'time-based boxcar' functions which might solve your problem. A time-based boxcar sum of 'window size' $\Delta t$ is defined at time $t$ to be the sum of all values between $t - \Delta t$ and $t$. This will be subject to discontinuities which you may or may not want. If you want older values to be downw... | null | CC BY-SA 2.5 | null | 2010-09-16T17:21:56.977 | 2010-09-24T04:24:59.887 | 2010-09-24T04:24:59.887 | 795 | 795 | null |
2738 | 2 | null | 2728 | 3 | null | You may find an [answer to a similar question](https://stats.stackexchange.com/questions/723/how-can-i-test-whether-my-clustering-of-binary-data-is-significant/749#749) useful. I have also used clValid but, as I recall, it was rather slow (at least for relatively large datasets).
| null | CC BY-SA 2.5 | null | 2010-09-16T18:21:20.133 | 2010-09-16T18:21:20.133 | 2017-04-13T12:44:24.947 | -1 | 339 | null |
2739 | 1 | 2741 | null | 4 | 1583 | Can someone give a concise, layman's explanation of an "If-Then" rule (as in rule-based systems). I am finding this term used frequently without anyone really defining it properly.
| What is an "If-Then" rule? | CC BY-SA 2.5 | null | 2010-09-16T18:36:13.733 | 2016-02-24T19:45:40.013 | null | null | 1121 | [
"machine-learning"
] |
2740 | 2 | null | 2684 | 2 | null | Higher correlation within subject gets you more power when the test being done is a differencing, equivalent to a paired t-test. The standard deviation used in calculating effect size is multiplied by $\sqrt{1-\rho}$. The standard deviation for difference scores (for a one-sample test) is $SD\sqrt{2-2\rho}$. This al... | null | CC BY-SA 3.0 | null | 2010-09-16T18:41:33.953 | 2016-03-04T05:59:37.200 | 2016-03-04T05:59:37.200 | 101484 | 1324 | null |
2741 | 2 | null | 2739 | 8 | null | It is just a simple classifier; so simple that it is better to explain it by example. Let's say you have a 3 class classification problem and information system with 4 continous predictors $X1,\ldots, X4$. Now you can define simple rules like:
```
IF X1<3.45 THEN Class1
IF X3>7.2 THEN Class2
IF X2<2.11 THEN Class1
IF X... | null | CC BY-SA 2.5 | null | 2010-09-16T19:20:07.583 | 2010-09-16T19:27:07.387 | 2010-09-16T19:27:07.387 | null | null | null |
2742 | 1 | 2747 | null | 11 | 14047 | Several statistical packages, such as SAS, SPSS, and R, allow you to perform some kind of factor rotation following a PCA.
- Why is a rotation necessary after a PCA?
- Why would you apply an oblique rotation after a PCA given that the aim of PCA is to produce orthogonal dimensions?
| On the use of oblique rotation after PCA | CC BY-SA 2.5 | null | 2010-09-16T19:28:10.130 | 2015-10-20T11:23:49.137 | 2015-10-20T11:23:49.137 | 28666 | 1154 | [
"pca",
"factor-analysis",
"factor-rotation"
] |
2743 | 1 | null | null | 3 | 3323 | I am attempting to calculate Cohen's d and then pool those estimates into a summary effect size. Can anyone help? (Stata or SPSS software owned).
| How to use Stata to pool Cohen's d? | CC BY-SA 3.0 | 0 | 2010-09-16T19:43:55.820 | 2019-07-18T19:46:54.897 | 2013-07-20T22:58:44.117 | 22047 | null | [
"spss",
"meta-analysis",
"stata",
"effect-size",
"cohens-d"
] |
2745 | 2 | null | 2742 | 4 | null | The problem with orthogonal dimensions is that the components can be uninterpretable. Thus, while oblique rotation (i.e., nonorthogonal dimensions) is technically less satisfying such a rotation sometimes enhances interpretablity of the resulting components.
| null | CC BY-SA 2.5 | null | 2010-09-16T19:58:55.770 | 2010-09-16T19:58:55.770 | null | null | null | null |
2746 | 1 | 2750 | null | 43 | 15847 | I would like to be able to efficiently generate positive-semidefinite (PSD) correlation matrices. My method slows down dramatically as I increase the size of matrices to be generated.
- Could you suggest any efficient solutions? If you are aware of any examples in Matlab, I would be very thankful.
- When generating a... | How to efficiently generate random positive-semidefinite correlation matrices? | CC BY-SA 3.0 | null | 2010-09-16T20:39:00.603 | 2021-01-30T16:41:25.453 | 2015-11-24T10:35:11.873 | 28666 | 1250 | [
"random-generation",
"correlation-matrix"
] |
2747 | 2 | null | 2742 | 9 | null | I think there are different opinions or views about PCA, but basically we often think of it as either a reduction technique (you reduce your features space to a smaller one, often much more "readable" providing you take care of properly centering/standardizing the data when it is needed) or a way to construct latent fa... | null | CC BY-SA 2.5 | null | 2010-09-16T20:45:22.383 | 2010-09-16T21:05:45.597 | 2017-04-13T12:44:24.947 | -1 | 930 | null |
2748 | 1 | null | null | 4 | 16696 | I am attempting to find a program that will let me conduct Cox regression on my matched case-control dataset.
Please assist.
p.s. I have STATA, SPSS, and MedCalc
| How to conduct conditional Cox regression for matched case-control study? | CC BY-SA 2.5 | null | 2010-09-16T20:46:35.240 | 2016-02-21T16:40:15.307 | 2010-09-18T09:50:10.163 | 521 | null | [
"regression",
"survival",
"frailty"
] |
2749 | 2 | null | 2748 | 2 | null | If you are using Stata, you can just look at the `stcox` command. Examples are available from [Stata](http://www.stata.com/capabilities/survivalsession.html) or [UCLA](http://www.ats.ucla.edu/stat/stata/topics/survival.htm) website.
Also, take a look at [Analysis of matched cohort data](http://www.stata-journal.com/sjp... | null | CC BY-SA 2.5 | null | 2010-09-16T20:56:40.193 | 2010-09-16T20:56:40.193 | null | null | 930 | null |
2750 | 2 | null | 2746 | 18 | null | You can do it backward: every matrix $C \in \mathbb{R}_{++}^p$ (the set of all symmetric $p \times p$ PSD matrices) can be decomposed as
$C=O^{T}DO$ where $O$ is an orthonormal matrix
To get $O$, first generate a random basis $(v_1,...,v_p)$ (where $v_i$ are random vectors, typically in $(-1,1)$). From there, use th... | null | CC BY-SA 3.0 | null | 2010-09-16T21:05:15.777 | 2015-11-03T11:21:36.323 | 2015-11-03T11:21:36.323 | 603 | 603 | null |
2751 | 2 | null | 2743 | 3 | null | In Stata:
1) Download the user-written package -[metan](http://ideas.repec.org/c/boc/bocode/s456798.html)- from the [SSC software library](http://ideas.repec.org/s/boc/bocode.html):
`. ssc install metan`
2) Assuming you have numbers, means and SDs for two groups, the syntax is of the form:
`. metan n1 mean1 sd1 n0 mean... | null | CC BY-SA 2.5 | null | 2010-09-16T21:09:15.640 | 2010-09-16T21:09:15.640 | null | null | 449 | null |
2752 | 2 | null | 2743 | 3 | null | For SPSS, look at :
- www.mathkb.com/Uwe/Forum.aspx/stat-consult/1201/Effect-Size-in-SPSS
- www.spsstools.net/Syntax/T-Test/StandardizedEffectsSize.txt (maybe better organized)
For Stata, I used [SIZEFX: Stata module to compute effect size correlations](http://ideas.repec.org/c/boc/bocode/s456738.html) (`findit si... | null | CC BY-SA 2.5 | null | 2010-09-16T21:10:09.010 | 2010-09-16T21:10:09.010 | null | null | 930 | null |
2753 | 2 | null | 2715 | 22 | null | Question your data. In the modern era of cheap RAM, we often work on large amounts of data. One 'fat-finger' error or 'lost decimal place' can easily dominate an analysis. Without some basic sanity checking, (or plotting the data, as suggested by others here) one can waste a lot of time. This also suggests using some b... | null | CC BY-SA 2.5 | null | 2010-09-16T21:32:16.190 | 2010-09-16T21:32:16.190 | null | null | 795 | null |
2755 | 2 | null | 2715 | 13 | null | For data organization/management, ensure that when you generate new variables in the dataset (for example, calculating body mass index from height and weight), the original variables are never deleted. A non-destructive approach is best from a reproducibility perspective. You never know when you might mis-enter a comma... | null | CC BY-SA 2.5 | null | 2010-09-16T22:36:18.300 | 2010-09-16T22:36:18.300 | null | null | 561 | null |
2756 | 2 | null | 411 | 3 | null | I can't give you additional reasons to use the Kolmogorov-Smirnov test. But, I can give you an important reason not to use it. It does not fit the tail of the distribution well. In this regard, a superior distribution fitting test is Anderson-Darling. As a second best, the Chi Square test is pretty good. Both are ... | null | CC BY-SA 2.5 | null | 2010-09-16T22:38:46.897 | 2010-09-16T22:38:46.897 | null | null | 1329 | null |
2757 | 2 | null | 2715 | 21 | null | Use software that shows the chain of programming logic from the raw data through to the final analyses/results. Avoid software like Excel where one user can make an undetectable error in one cell, that only manual checking will pick up.
| null | CC BY-SA 2.5 | null | 2010-09-16T22:39:16.767 | 2010-09-16T22:39:16.767 | null | null | null | null |
2761 | 2 | null | 305 | 1 | null | I would take the opposite view here. Why bother with the Welch test when the standard unpaired student t test gives you nearly identical results. I studied this issue a while back and I explored a range of scenarios in an attempt to break down the t test and favor the Welch test. To do so I used sample sizes up to 5... | null | CC BY-SA 2.5 | null | 2010-09-16T23:52:59.773 | 2010-09-16T23:52:59.773 | null | null | 1329 | null |
2763 | 2 | null | 2746 | 16 | null | An even simpler characterization is that for real matrix $A$, $A^TA$ is positive semidefinite. To see why this is the case, one only has to prove that $y^T (A^TA) y \ge 0$ for all vectors $y$ (of the right size, of course). This is trivial: $y^T (A^TA) y = (Ay)^T Ay = ||Ay||$ which is nonnegative. So in Matlab, simply ... | null | CC BY-SA 2.5 | null | 2010-09-17T00:14:46.137 | 2010-09-17T00:14:46.137 | null | null | 795 | null |
2764 | 1 | null | null | 2 | 1663 | I am attempting to calculate Orwin's (1983) modification of Rosenthal's (1979) Fail-safe N for my meta-analysis of Odds Ratios.
However, all the equations I am finding are using Cohen's d, which I cannot calculate (I don't have two groups).
I have STATA, SPSS, and MedCalc at my disposal.
| Calculating Orwin's (1983) modified Fail-safe N in a meta-analysis with Odds Ratio as summary statistic? | CC BY-SA 2.5 | null | 2010-09-17T00:30:13.537 | 2010-09-17T20:18:34.320 | 2010-09-17T20:18:34.320 | null | null | [
"meta-analysis",
"cohens-d",
"odds-ratio"
] |
2765 | 2 | null | 2730 | 6 | null | In my line of work, I've found this one to be quite useful: [Statistical Methods for Psychology (Howell, 2009)](http://rads.stackoverflow.com/amzn/click/0495597848)
| null | CC BY-SA 2.5 | null | 2010-09-17T00:41:03.327 | 2010-09-17T00:41:03.327 | null | null | 1334 | null |
2766 | 2 | null | 2639 | 0 | null | It's not clear why you must have a confidence interval. As @whuber pointed out, there are better ways to compare distributions. You are losing some information by looking only at the mean. However, if you must, you might want to compute a confidence interval on `k1_mean` based on the 40000 observations. See [Wikipedia]... | null | CC BY-SA 2.5 | null | 2010-09-17T02:40:40.300 | 2010-09-17T02:40:40.300 | null | null | 795 | null |
2767 | 2 | null | 2516 | 24 | null | Hypothesis testing traditionally focused on p values to derive statistical significance when alpha is less than 0.05 has a major weakness. And, that is that with a large enough sample size any experiment can eventually reject the null hypothesis and detect trivially small differences that turn out to be statistically ... | null | CC BY-SA 2.5 | null | 2010-09-17T04:11:52.573 | 2010-09-17T04:11:52.573 | null | null | 1329 | null |
2768 | 1 | null | null | 11 | 15243 | I was asked such a question as "Did you do any consistency check in your daily work?" during a phone interview for a Biostatistician position. I don't know what to answer. Any information is appreciated.
| What is a consistency check? | CC BY-SA 2.5 | null | 2010-09-17T04:36:00.473 | 2022-01-20T05:32:41.387 | 2010-09-17T11:44:38.123 | 159 | 1336 | [
"validation"
] |
2769 | 2 | null | 2742 | 4 | null | Basic Points
- Rotation can make interpretation of components clearer
- Oblique rotation often makes more theoretical sense. I.e., Observed variables can be explained in terms of a smaller number of correlated components.
Example
- 10 tests all measuring ability with some measuring verbal and some measuring spatia... | null | CC BY-SA 2.5 | null | 2010-09-17T05:01:12.770 | 2010-09-17T05:01:12.770 | null | null | 183 | null |
2770 | 1 | 2773 | null | 13 | 2716 | I'm an undergraduate statistics student looking for a good treatment of clinical trials analysis. The text should cover the fundamentals of experimental design, blocking, power analysis, latin squares design, and cluster randomization designs, among other topics.
I have an undergraduate knowledge of mathematical stati... | Good text on Clinical Trials? | CC BY-SA 2.5 | null | 2010-09-17T05:05:49.097 | 2022-12-06T02:54:37.377 | 2010-09-17T06:36:30.523 | 183 | 1118 | [
"references",
"teaching",
"clinical-trials"
] |
2771 | 2 | null | 2516 | 5 | null | The short answer is "no". Research on hypothesis testing in the asymptotic regime of infinite observations and multiple hypotheses has been very, very active in the past 15-20 years, because of microarray data and financial data applications. The long answer is in the course page of Stat 329, "Large-Scale Simultaneous ... | null | CC BY-SA 4.0 | null | 2010-09-17T05:49:34.810 | 2021-10-15T14:16:05.627 | 2021-10-15T14:16:05.627 | 321901 | 30 | null |
2772 | 1 | null | null | 23 | 5444 | In two papers in [1986](http://dx.doi.org/10.1016/0304-405X%2886%2990027-9) and [1988](http://dx.doi.org/10.1016/0304-405X%2888%2990062-1), Connor and Korajczyk proposed an approach to modeling asset returns. Since these time series have usually more assets than time period observations, they proposed to perform a PCA ... | What is the difference between PCA and asymptotic PCA? | CC BY-SA 3.0 | null | 2010-09-17T06:11:36.290 | 2022-10-01T13:29:06.397 | 2011-06-19T09:20:09.777 | null | 30 | [
"pca",
"econometrics"
] |
2773 | 2 | null | 2770 | 8 | null | I would definitively recommend [Design and Analysis of Clinical Trials: Concepts and Methodologies](http://books.google.fr/books?id=HXMUEjZ4vyAC&dq=Design+and+Analysis+of+Clinical+Trials:+Concepts+and+Methodologies&printsec=frontcover&source=bn&hl=fr&ei=YQuTTPG-KseLswbt9JH5CQ&sa=X&oi=book_result&ct=result&resnum=4&ved=... | null | CC BY-SA 4.0 | null | 2010-09-17T06:32:58.730 | 2022-12-06T02:53:13.523 | 2022-12-06T02:53:13.523 | 362671 | 930 | null |
2774 | 2 | null | 2768 | 8 | null | I suppose this has to do with some form of Quality Control about data integrity, and more specifically that you regularly check that your working database isn't corrupted (due to error during transfer, copy, or after an update or a sanity check). This may also mean ensuring that your intermediate computation are double... | null | CC BY-SA 2.5 | null | 2010-09-17T06:43:26.070 | 2010-09-17T06:49:06.947 | 2010-09-17T06:49:06.947 | 930 | 930 | null |
2775 | 2 | null | 2746 | 3 | null | The simplest method is the one above, which is a simulation of a random dataset and the computation of the [Gramian](http://en.wikipedia.org/wiki/Gramian_matrix). A word of caution: The resulting matrix will not be uniformly random, in that its decomposition, say $U^TSU$ will have rotations not distributed according to... | null | CC BY-SA 2.5 | null | 2010-09-17T06:52:22.090 | 2010-09-17T14:39:00.353 | 2010-09-17T14:39:00.353 | 30 | 30 | null |
2776 | 2 | null | 2764 | 1 | null | So, perhaps check these additional resources: [http://j.mp/d8znoP](http://j.mp/d8znoP) for SPSS. Don't know about Stata.
There is some R code about fail-safe N in the following handout: [Tests for funnel plot asymmetry and failsafe N](http://www.biostat.uzh.ch/teaching/phd/doktorandenseminar/schroedle.pdf), but I didn'... | null | CC BY-SA 2.5 | null | 2010-09-17T07:10:43.227 | 2010-09-17T07:10:43.227 | null | null | 930 | null |
2777 | 1 | 2781 | null | 34 | 9555 | Context:
Imagine you had a longitudinal study which measured a dependent variable (DV) once a week for 20 weeks on 200 participants. Although I'm interested in general, typical DVs that I'm thinking of include job performance following hire or various well-being measures following a clinical psychology intervention.
I ... | Modelling longitudinal data where the effect of time varies in functional form between individuals | CC BY-SA 2.5 | null | 2010-09-17T07:12:27.057 | 2014-09-19T01:29:53.793 | null | null | 183 | [
"repeated-measures",
"random-effects-model",
"latent-class"
] |
2778 | 2 | null | 2730 | 11 | null | The classics I think are Winer and Kirk, both cover essentially only ANOVA and ANCOVA. You can probably get used copys for cheap (e.g., I own a Winer second edition from 71 bought via AMAZON for less than 10$):
[Winer - Statistical Principles In Experimental Design](http://rads.stackoverflow.com/amzn/click/0070709823)
... | null | CC BY-SA 2.5 | null | 2010-09-17T08:43:52.880 | 2010-09-17T08:43:52.880 | null | null | 442 | null |
2779 | 2 | null | 2711 | 1 | null | For the problem above, I have used a really simple metric.
I wanted to asses how good is my hit!
If I have, for example, errors 4, 2, 0.002, 3, 6 then I choose h1=0.002 as hit, and h2=2 as the closest error.
X/h1 + X/h2 = 100% => X/h1 and X/h2 is %.
| null | CC BY-SA 2.5 | null | 2010-09-17T08:44:09.753 | 2010-09-17T08:44:09.753 | null | null | 1313 | null |
2780 | 2 | null | 2764 | 2 | null | Normally you always find ways to convert effect size. For example you can calculate $r$ from $d$ and back. So you surely will be able to converts odds ratio to Cohen's $d$.
One book I usually found a good resource for stuff like that is the one by Rosenthal & Rosnow. I think it was this one:
[Rosenthal & Rosnow - Essen... | null | CC BY-SA 2.5 | null | 2010-09-17T08:57:59.967 | 2010-09-17T08:57:59.967 | null | null | 442 | null |
2781 | 2 | null | 2777 | 21 | null | I would suggest to look at the following three directions:
- longitudinal clustering: this is unsupervised, but you use k-means approach relying on the Calinsky criterion for assessing quality of the partitioning (package kml, and references included in the online help); so basically, it won't help identifying specifi... | null | CC BY-SA 2.5 | null | 2010-09-17T08:59:36.203 | 2010-09-17T13:52:39.000 | 2010-09-17T13:52:39.000 | 930 | 930 | null |
2782 | 2 | null | 2715 | 28 | null | If you're deciding between two ways of analysing your data, try it both ways and see if it makes a difference.
This is useful in many contexts:
- To transform or not transform
- Non-parametric or parameteric test
- Spearman's or Pearson's correlation
- PCA or factor analysis
- Whether to use the arithmetic mean o... | null | CC BY-SA 2.5 | null | 2010-09-17T09:40:03.130 | 2010-09-19T10:13:15.170 | 2010-09-19T10:13:15.170 | 183 | 183 | null |
2783 | 2 | null | 2746 | 4 | null | You haven't specified a distribution for the matrices. Two common ones are the Wishart and inverse Wishart distributions. The [Bartlett decomposition](http://en.wikipedia.org/wiki/Wishart_distribution#Bartlett_decomposition) gives a Cholesky factorisation of a random Wishart matrix (which can also be efficiently solved... | null | CC BY-SA 2.5 | null | 2010-09-17T11:10:21.900 | 2010-09-17T11:10:21.900 | null | null | 495 | null |
2784 | 2 | null | 2061 | 3 | null | In the category (3) of humorous videos, check out ['Statz rappers'](http://video.google.com/videoplay?docid=489221653835413043#); general interest. (Pretty funny even to older people ;-).)
| null | CC BY-SA 2.5 | null | 2010-09-17T11:11:41.793 | 2010-09-17T11:11:41.793 | null | null | 919 | null |
2785 | 2 | null | 2768 | 18 | null | To chl's list, which focuses on frank data processing errors, I would add checks for subtler errors to address the following questions and issues (given in no particular order and certainly incomplete):
- Assuming database integrity, are the data reasonable? Do they roughly conform with expectations or conventional m... | null | CC BY-SA 3.0 | null | 2010-09-17T11:41:19.193 | 2013-07-29T05:58:25.537 | 2013-07-29T05:58:25.537 | 7290 | 919 | null |
2786 | 2 | null | 2746 | 11 | null | As a variation on kwak's answer: generate a diagonal matrix $\mathbf{D}$ with random nonnegative eigenvalues from a distribution of your choice, and then perform a similarity transformation $\mathbf{A}=\mathbf{Q}\mathbf{D}\mathbf{Q}^T$ with $\mathbf{Q}$ a [Haar-distributed pseudorandom orthogonal matrix](http://dx.doi.... | null | CC BY-SA 2.5 | null | 2010-09-17T12:12:34.027 | 2010-09-17T12:12:34.027 | null | null | 830 | null |
2787 | 1 | 2790 | null | 1 | 163 | It was recommended to me on the [User Interface Stack Exchange](https://ux.stackexchange.com/questions/1543/whats-recommended-for-the-number-of-users-to-use-in-qualitative-and-quantitative) that I also ask this question here. So....
I'm currently rebuilding our entire intranet from scratch, mostly because the tech behi... | What's recommended for the number of users to use in qualitative and quantitative usability studies? | CC BY-SA 2.5 | null | 2010-09-17T12:52:07.833 | 2011-02-23T01:02:50.827 | 2017-04-12T07:29:42.590 | -1 | 593 | [
"sample-size",
"hypothesis-testing"
] |
2788 | 1 | 2789 | null | 7 | 2045 | I hope you all won't mind a basic question.
We are examining the effects of a compound at various concentration on the behaviour of an organism. The compound is administered once at the beginning of the time course. Observations are made every minute for a period of 2 hours. Each concentration is applied to 12 indiv... | Combining repeated experiments into one dataset | CC BY-SA 3.0 | null | 2010-09-17T12:54:10.173 | 2012-04-19T14:03:34.913 | 2012-04-19T14:03:34.913 | 10322 | 1079 | [
"multilevel-analysis"
] |
2789 | 2 | null | 2788 | 8 | null | Just add "experiment" as an effect to your model, that should account for the shift between experiments and let you gain the power of increased N across experiments to detect effects of concentration and time.
In R, if using ANOVA and treating time as a factor (e.g. not numeric), then do:
```
library(ez)
ezANOVA(
... | null | CC BY-SA 2.5 | null | 2010-09-17T13:14:32.587 | 2010-09-17T13:46:22.303 | 2017-04-13T12:44:55.360 | -1 | 364 | null |
2790 | 2 | null | 2787 | 3 | null | Not an exact answer to your question (which is kind of not understandable enough for conrete answers, see Srikant's comment), but a more general one:
If you want to do statistical testing on your data it is advisable to make sure that your sample conforms to the [central limit theorem](http://en.wikipedia.org/wiki/Cent... | null | CC BY-SA 2.5 | null | 2010-09-17T13:27:27.217 | 2010-09-17T13:27:27.217 | 2017-04-13T12:44:55.630 | -1 | 442 | null |
2791 | 2 | null | 2770 | 4 | null | [Experimental Design for the Life Sciences](http://rads.stackoverflow.com/amzn/click/019928511X), by Ruxton & Colegrave, is a nice book and is aimed primarily at undergraduates.
| null | CC BY-SA 2.5 | null | 2010-09-17T13:36:22.340 | 2010-09-17T13:36:22.340 | null | null | 266 | null |
2792 | 2 | null | 2788 | 6 | null | I think a way in which you could analyze the two experiments together is by defining a [multilevel/hierarchical model](http://en.wikipedia.org/wiki/Hierarchical_linear_modeling). The individuals are nested within each experiment.
The standard for this approach is [Gelman's book](http://rads.stackoverflow.com/amzn/click... | null | CC BY-SA 2.5 | null | 2010-09-17T13:39:43.130 | 2010-09-17T13:50:02.507 | 2010-09-17T13:50:02.507 | 442 | 442 | null |
2793 | 2 | null | 2516 | 13 | null | One simple point not made directly in another answer is that it's simply not true that "all null hypotheses are false."
The simple hypothesis that a physical coin has heads probability exactly equal to 0.5, ok, that is false.
But the compound hypothesis that a physical coin has heads probability greater than 0.499 and ... | null | CC BY-SA 2.5 | null | 2010-09-17T15:26:45.237 | 2010-09-17T15:26:45.237 | null | null | 1122 | null |
2794 | 1 | 2798 | null | 6 | 5009 | I know the significance of the slope of a LMS linear regression can be calculated using the r2 coefficient of determination and looking up the appropriate value in an F table. However, I was thinking of perhaps making this "more robust" by replacing the LMS linear regression with a repeated median straight line fit, an... | Significance of the slope of a straight line fit | CC BY-SA 2.5 | null | 2010-09-17T15:44:08.690 | 2014-05-05T09:31:44.037 | null | null | 226 | [
"regression",
"statistical-significance",
"robust"
] |
2796 | 2 | null | 1621 | 4 | null | The most well know argument is that the fisher metric, being invariant to coordinate transforms, can be used to formulate an uninformed prior (Jeffreys prior). Not sure I buy it!
Less well known, is that sometimes these "integrated quantities" turn out to be divergences and such, one may argue that the fisher distances... | null | CC BY-SA 2.5 | null | 2010-09-17T16:31:46.073 | 2010-09-17T16:31:46.073 | null | null | 1342 | null |
2797 | 2 | null | 2245 | 9 | null | In addition to the excellent answer above, there is a statistical method that can get you closer to demonstrating causality. It is Granger Causality that demonstrates that one independent variable occurring before a dependent variable has a causal effect or not. I introduce this method in an easy to follow presentati... | null | CC BY-SA 2.5 | null | 2010-09-17T16:49:17.907 | 2010-09-17T16:49:17.907 | null | null | 1329 | null |
2798 | 2 | null | 2794 | 3 | null | No, F tests are based on the assumption that lowest sum of residual squares is optimal. It does not hold in case of robust regression, where the criterion is different.
For instance, effectively one may consider robust regression as least squares on data stripped from outliers; using $r^2$ on all data in this case adds... | null | CC BY-SA 2.5 | null | 2010-09-17T17:37:00.617 | 2010-09-17T17:37:00.617 | null | null | null | null |
2801 | 2 | null | 2794 | 3 | null | No need to reinvent the wheel. There is an alternative, robust, R^2 measure with very good statistical properties:
[A robust coefficient of determination for regression, O Renauda](http://www.google.com/url?sa=t&source=web&cd=3&ved=0CBsQFjAC&url=http%3A%2F%2Fwww.hec.unige.ch%2Fwww%2Fdms%2Fhec_en%2Fvictoriafeser%2Freche... | null | CC BY-SA 2.5 | null | 2010-09-17T18:36:17.920 | 2010-09-21T16:43:01.597 | 2010-09-21T16:43:01.597 | 603 | 603 | null |
2802 | 2 | null | 2794 | 1 | null | I would simply use the standard regression output to evaluate the significance of the slope coefficient. I mean by that looking at the coefficient itself, its standard error, t stat (# of standard errors = Coefficient/Standard error), p value, and confidence interval. The p value directly addresses the statistical si... | null | CC BY-SA 2.5 | null | 2010-09-17T18:57:23.110 | 2010-09-17T18:57:23.110 | null | null | 1329 | null |
2805 | 2 | null | 2077 | 13 | null | I still think using the % change from one period to the next is the best way to render a non-stationary variable stationary as you first suggest. A transformation such as a log works reasonably well (it flattens the non-stationary quality; but does not eliminate it entirely).
The third way is to deseasonalize and de... | null | CC BY-SA 2.5 | null | 2010-09-17T21:44:07.557 | 2010-09-17T21:44:07.557 | null | null | 1329 | null |
2806 | 1 | null | null | 55 | 32217 | I previously asked this on StackOverflow, but it seems like it might be more appropriate here, given that it didn't get any answers on SO. It's kind of at the intersection between statistics and programming.
I need to write some code to do PCA (Principal Component Analysis). I've browsed through the well-known algorit... | Best PCA algorithm for huge number of features (>10K)? | CC BY-SA 3.0 | null | 2010-09-18T02:08:24.590 | 2019-04-25T07:42:50.480 | 2015-06-09T23:33:35.577 | 7291 | 1347 | [
"pca",
"algorithms",
"model-evaluation",
"high-dimensional"
] |
2807 | 2 | null | 2492 | 68 | null | IMHO normality tests are absolutely useless for the following reasons:
- On small samples, there's a good chance that the true distribution of the population is substantially non-normal, but the normality test isn't powerful to pick it up.
- On large samples, things like the T-test and ANOVA are pretty robust to non-... | null | CC BY-SA 2.5 | null | 2010-09-18T02:32:42.093 | 2010-09-18T02:32:42.093 | null | null | 1347 | null |
2808 | 2 | null | 1761 | 3 | null | GD047 linked to a UC Berkeley video on statistics, and I had another one. It's similar to MIT's Open Courseware (maybe a little lower quality). This one is for the Intro Statistics and Probability Class. It unfortunately only has 9 episodes, but it's free :).
[UC Berkeley STATS 20 Webcast](http://webcast.berkeley.edu/c... | null | CC BY-SA 2.5 | null | 2010-09-18T02:43:41.650 | 2010-09-18T02:43:41.650 | null | null | 1118 | null |
2809 | 2 | null | 2777 | 8 | null | I'd recommend taking a look at a couple of papers by Heping Zhang using adaptive splines for modeling longitudinal data:
- Multivariate adaptive splines for analysis of longitudinal data (Free PDF)
- Mixed effects multivariate adaptive splines model for the analysis of longitudinal and growth curve data
In addition... | null | CC BY-SA 2.5 | null | 2010-09-18T02:56:41.683 | 2010-09-19T09:58:00.853 | 2010-09-19T09:58:00.853 | 183 | 251 | null |
2810 | 2 | null | 2806 | 1 | null | See Sam Roweis' paper, [EM Algorithms for PCA and SPCA](http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.16.3330).
| null | CC BY-SA 2.5 | null | 2010-09-18T03:04:07.327 | 2010-09-18T03:04:07.327 | null | null | 251 | null |
2811 | 2 | null | 2777 | 6 | null | It looks to me like [Growth Mixture Models](http://dx.doi.org/10.1007/s10940-007-9036-0) might have potential to allow you to examine your error variance. ([PDF](http://statmodel2.com/download/kreutermuthen2006_34.pdf) here). (I'm not sure what multiplicative heteroscedastic models are, but I will definitely have to ch... | null | CC BY-SA 2.5 | null | 2010-09-18T03:45:56.633 | 2010-09-18T03:45:56.633 | null | null | 1036 | null |
2812 | 2 | null | 2806 | 4 | null | It sounds like maybe you want to use the [Lanczos Algorithm](http://en.wikipedia.org/wiki/Lanczos_iteration). Failing that, you might want to consult [Golub & Van Loan.](http://rads.stackoverflow.com/amzn/click/0801854148) I once coded a SVD algorithm (in SML, of all languages) from their text, and it worked reasonably... | null | CC BY-SA 2.5 | null | 2010-09-18T04:08:12.430 | 2010-09-18T04:08:12.430 | null | null | 795 | null |
2813 | 2 | null | 2806 | 3 | null | I'd suggest trying [kernel PCA](http://en.wikipedia.org/wiki/Kernel_PCA) which has a time/space complexity dependent on the number of examples (N) rather than number of features (P), which I think would be more suitable in your setting (P>>N)). Kernel PCA basically works with NxN kernel matrix (matrix of similarities b... | null | CC BY-SA 2.5 | null | 2010-09-18T04:25:12.827 | 2010-09-18T04:25:12.827 | null | null | 881 | null |
2814 | 2 | null | 2615 | 2 | null | The proof should be rather simple. Let $C$ be your correlation matrix (it has all ones on the diagonal). You multiply each element on the off diagonal by $(1+k)$. This is equivalent to computing the matrix $\hat{C} = (1+k) C - k \operatorname{diag}(C) = (1+k)C - k I,$ where $\operatorname{diag}$ is the diagonal part of... | null | CC BY-SA 2.5 | null | 2010-09-18T05:03:55.393 | 2010-09-18T20:55:20.910 | 2010-09-18T20:55:20.910 | 795 | 795 | null |
2815 | 2 | null | 2679 | 0 | null | you could model the traffic flow of red cars as a compound poisson distribution.
suppose we look at the number N of cars that pass in a fixed time [8 hours, perhaps].
N should have a poisson distribution, according to your assumptions. let $\lambda$ be the poisson parameter for N.
suppose all of the cars are colorless... | null | CC BY-SA 2.5 | null | 2010-09-18T05:07:02.720 | 2010-09-18T05:07:02.720 | null | null | 1112 | null |
2816 | 2 | null | 2770 | 5 | null | It's difficult to go past Piantadosi's [Clinical Trials: A Methodologic Perspective](https://rads.stackoverflow.com/amzn/click/com/0471727814) at least as a starting point. It is quite comprehensive and covers the topics you want and much more - including ethics, history, reporting, fraud, meta-analyses, randomisation,... | null | CC BY-SA 4.0 | null | 2010-09-18T08:51:04.823 | 2022-12-06T00:45:33.433 | 2022-12-06T00:45:33.433 | 11887 | 521 | null |
2817 | 2 | null | 2806 | 2 | null | I seem to recall that it is possible to perform PCA by computing the eigen-decomposition of X^TX rather than XX^T and then transform to get the PCs. However I can't remember the details off-hand, but it is in Jolliffe's (excellent) book and I'll look it up when I am next at work. I'd transliterate the linear algebra ... | null | CC BY-SA 2.5 | null | 2010-09-18T09:26:40.507 | 2010-09-18T09:26:40.507 | null | null | 887 | null |
2818 | 2 | null | 2748 | 4 | null | Half the battle with many questions is understanding the terminology. Matching implies within group (or within pair) correlation. Under appropriate circumstances matching can be dealt with paired t-tests, conditional logistic regression or mixed effects models.
In survival analysis (or time to event analysis), within ... | null | CC BY-SA 2.5 | null | 2010-09-18T09:49:20.960 | 2010-10-02T08:15:01.607 | 2010-10-02T08:15:01.607 | 521 | 521 | null |
2819 | 1 | 2822 | null | 30 | 14249 | I have p values from a lot of tests and would like to know whether there is actually something significant after correcting for multiple testing. The complication: my tests are not independent. The method I am thinking about (a variant of Fisher's Product Method, Zaykin et al., [Genet Epidemiol](http://www.ncbi.nlm.nih... | Correcting p values for multiple tests where tests are correlated (genetics) | CC BY-SA 2.5 | null | 2010-09-18T11:21:46.510 | 2012-11-05T10:09:29.377 | 2010-10-09T20:11:09.663 | 449 | 1352 | [
"correlation",
"multiple-comparisons",
"statistical-significance",
"genetics"
] |
2820 | 2 | null | 2819 | 3 | null | Using a method like bonferroni is fine, the problem is that if you have many tests you are not likely to find many "discoveries".
You can go with the FDR approach for dependent tests (see [here for details](http://en.wikipedia.org/wiki/False_discovery_rate#Dependent_tests)) the problem is that I am not sure if you can ... | null | CC BY-SA 2.5 | null | 2010-09-18T12:03:38.633 | 2010-09-18T12:03:38.633 | null | null | 253 | null |
2821 | 2 | null | 328 | 1 | null | Oren, it is useful to define what aspects of finance you intend to tackle. Statistics is a tool when seen from the econometrics perspective (in terms of assessing the plausibility of a proposed model / theory) or can be the first or primary line of attack when seen from the machine learning side - that is you go low on... | null | CC BY-SA 3.0 | null | 2010-09-18T12:21:00.450 | 2012-08-27T15:59:23.840 | 2012-08-27T15:59:23.840 | 7290 | 1349 | null |
2822 | 2 | null | 2819 | 35 | null | This is actually a hot topic in Genomewide analysis studies (GWAS)! I am not sure the method you are thinking of is the most appropriate in this context. Pooling of p-values was described by some authors, but in a different context (replication studies or meta-analysis, see e.g. (1) for a recent review). Combining SNP ... | null | CC BY-SA 2.5 | null | 2010-09-18T14:32:08.507 | 2010-10-06T11:23:22.413 | 2010-10-06T11:23:22.413 | 930 | 930 | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.