File size: 7,763 Bytes
0ab7b0c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 |
import pandas as pd
# import numpy as np
# from warnings import warn
# 2018.11.07 Created by Eamon.Zhang
# 2018.11.12 change into fit() transform() format
class GroupingRareValues():
"""
Grouping the observations that show rare labels into a unique category ('rare')
Parameters
----------
"""
def __init__(self, mapping=None, cols=None, threshold=0.01):
self.cols = cols
self.mapping = mapping
self._dim = None
self.threshold = threshold
def fit(self, X, y=None, **kwargs):
"""Fit encoder according to X and y.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Training vectors, where n_samples is the number of samples
and n_features is the number of features.
y : array-like, shape = [n_samples]
Target values.
Returns
-------
self : encoder
Returns self.
"""
self._dim = X.shape[1]
_, categories = self.grouping(
X,
mapping=self.mapping,
cols=self.cols,
threshold=self.threshold
)
self.mapping = categories
return self
def transform(self, X):
"""Perform the transformation to new categorical data.
Will use the mapping (if available) and the column list to encode the
data.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Returns
-------
X : Transformed values with encoding applied.
"""
if self._dim is None:
raise ValueError('Must train encoder before it can be used to transform data.')
# make sure that it is the right size
if X.shape[1] != self._dim:
raise ValueError('Unexpected input dimension %d, expected %d' % (X.shape[1], self._dim,))
X, _ = self.grouping(
X,
mapping=self.mapping,
cols=self.cols,
threshold=self.threshold
)
return X
def grouping(self, X_in, threshold, mapping=None, cols=None):
"""
Grouping the observations that show rare labels into a unique category ('rare')
"""
X = X_in.copy(deep=True)
# if cols is None:
# cols = X.columns.values
if mapping is not None: # transform
mapping_out = mapping
for i in mapping:
column = i.get('col') # get the column name
X[column] = X[column].map(i['mapping'])
# try:
# X[column] = X[column].astype(int)
# except ValueError as e:
# X[column] = X[column].astype(float)
else: # fit
mapping_out = []
for col in cols:
# if util.is_category(X[col].dtype):
# categories = X[col].cat.categories
# else:
temp_df = pd.Series(X[col].value_counts()/len(X))
mapping = { k: ('rare' if k not in temp_df[temp_df >= threshold].index else k)
for k in temp_df.index}
mapping = pd.Series(mapping)
mapping_out.append({'col': col, 'mapping': mapping, 'data_type': X[col].dtype}, )
return X, mapping_out
#==============================================================================
# def rare_imputation(X_train, X_test, variable):
#
# # find the most frequent category
# frequent_cat = X_train.groupby(variable)[variable].count().sort_values().tail(1).index.values[0]
#
# # find rare labels
# temp = X_train.groupby([variable])[variable].count()/np.float(len(X_train))
# rare_cat = [x for x in temp.loc[temp<0.05].index.values]
#
# # create new variables, with Rare labels imputed
#
# # by the most frequent category
# X_train[variable+'_freq_imp'] = np.where(X_train[variable].isin(rare_cat), frequent_cat, X_train[variable])
# X_test[variable+'_freq_imp'] = np.where(X_test[variable].isin(rare_cat), frequent_cat, X_test[variable])
#
# # by adding a new label 'Rare'
# X_train[variable+'_rare_imp'] = np.where(X_train[variable].isin(rare_cat), 'Rare', X_train[variable])
# X_test[variable+'_rare_imp'] = np.where(X_test[variable].isin(rare_cat), 'Rare', X_test[variable])
#==============================================================================
# 2018.11.26 created by Eamon.Zhang
class ModeImputation():
"""
Replacing the rare label by most frequent label
Parameters
----------
"""
def __init__(self, mapping=None, cols=None, threshold=0.01):
self.cols = cols
self.mapping = mapping
self._dim = None
self.threshold = threshold
def fit(self, X, y=None, **kwargs):
"""Fit encoder according to X and y.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Training vectors, where n_samples is the number of samples
and n_features is the number of features.
y : array-like, shape = [n_samples]
Target values.
Returns
-------
self : encoder
Returns self.
"""
self._dim = X.shape[1]
_, categories = self.impute_with_mode(
X,
mapping=self.mapping,
cols=self.cols,
threshold=self.threshold
)
self.mapping = categories
return self
def transform(self, X):
"""Perform the transformation to new categorical data.
Will use the mapping (if available) and the column list to encode the
data.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Returns
-------
X : Transformed values with encoding applied.
"""
if self._dim is None:
raise ValueError('Must train encoder before it can be used to transform data.')
# make sure that it is the right size
if X.shape[1] != self._dim:
raise ValueError('Unexpected input dimension %d, expected %d' % (X.shape[1], self._dim,))
X, _ = self.impute_with_mode(
X,
mapping=self.mapping,
cols=self.cols,
threshold=self.threshold
)
return X
def impute_with_mode(self, X_in, threshold, mapping=None, cols=None):
"""
Grouping the observations that show rare labels into a unique category ('rare')
"""
X = X_in.copy(deep=True)
# if cols is None:
# cols = X.columns.values
if mapping is not None: # transform
mapping_out = mapping
for i in mapping:
column = i.get('col') # get the column name
X[column] = X[column].map(i['mapping'])
# try:
# X[column] = X[column].astype(int)
# except ValueError as e:
# X[column] = X[column].astype(float)
else: # fit
mapping_out = []
for col in cols:
# if util.is_category(X[col].dtype):
# categories = X[col].cat.categories
# else:
temp_df = pd.Series(X[col].value_counts()/len(X))
median = X[col].mode()[0]
mapping = { k: (median if k not in temp_df[temp_df >= threshold].index else k)
for k in temp_df.index}
mapping = pd.Series(mapping)
mapping_out.append({'col': col, 'mapping': mapping, 'data_type': X[col].dtype}, )
return X, mapping_out
|