-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathFeature_Selection.py
181 lines (163 loc) · 7.08 KB
/
Feature_Selection.py
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
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.feature_selection import SelectKBest
from sklearn.feature_selection import chi2
from sklearn.linear_model import LogisticRegression
from sklearn.feature_selection import RFE
from sklearn.decomposition import PCA
from sklearn.ensemble import ExtraTreesClassifier
def plot_feature_importance1(features,importances,indices,feature_names):
plt.figure()
print(features.shape[1])
plt.title("Feature importances")
plt.bar(range(features.shape[1]), importances[indices], color="r", yerr=std[indices], align="center")
#plt.xticks(range(features.shape[1]), indices)
plt.xticks(range(features.shape[1]), feature_names)
plt.xlim([-1, features.shape[1]])
plt.show()
return
def plot_feature_importance2(features,importances,indices,feature_names, category_name,num_max_features):
f, ax = plt.subplots(figsize=(11, 9))
plt.title("Feature ranking for "+category_name, fontsize=20)
print("range(features.shape[1]) "+str(range(features.shape[1])))
#plt.bar(features.shape[1], importances[indices], color="b", align="center")
plt.bar(range(num_max_features), importances[indices[range(num_max_features)]], color="b", align="center")
#feature_names = df.columns # e.g. ['A', 'B', 'C', 'D', 'E']
plt.xticks(range(num_max_features), feature_names,rotation='vertical')# plot features by names
#plt.xticks(range(features.shape[1]), indices) plot feature by indices
#plt.xlim([-1, features.shape[1]])
plt.xlim([-1, num_max_features])
plt.ylabel("importance", fontsize=18)
plt.xlabel("index of the feature", fontsize=18)
plt.show()
return
def Extract_Features_From_File(file_path):
ranks = []
features_vectors = []
with open(file_path, 'r') as filep:
for item in filep:
row = item.split(' ')
rank = row[0]
ranks.append(int(rank))
vector = []
for i in range(2, len(row)):
feat = row[i].split(':')
if len(feat)>1:
vector.append(float(feat[1]))
features_vectors.append(vector)
return features_vectors,ranks
def Feature_Selection_Using_UniVariate(X,Y):
# feature extraction using Univariate Selection
print("Feature extraction using Univariate Selection")
test = SelectKBest(score_func=chi2, k=4)
fit = test.fit(X, Y)
# summarize scores
np.set_printoptions(precision=3)
print(fit.scores_)
features = fit.transform(X)
print(features.shape)
# summarize selected features
print(features[0:5, :])
return
def Feature_Selection_Using_Recursive_Feature_Elimination(X,Y):
# Feature Extraction with RFE
print("Feature Extraction with RFE")
model = LogisticRegression()
num_requested_features = int(X.shape[1]/2)
print("num_requested_features "+str(num_requested_features))
rfe = RFE(model, 3)
print("Fitting")
fit = rfe.fit(X, Y)
print("Num Features: " + str(fit.n_features_))
print("Selected Features " + str(fit.support_))
print("Feature Ranking: " + str(fit.ranking_))
return
def Feature_Selection_Using_Principal_Component_Analysis(X,Y):
# feature extraction
print("Principal Component Analysis")
pca = PCA(n_components=3)
fit = pca.fit(X)
# summarize components
print("Explained Variance: "+str(fit.explained_variance_ratio_))
print(fit.components_)
return
def get_feature_names(num_time_periods):
feature_names = []
for i in range(num_time_periods):
feature_names.append("#Rating_1_P"+str(i+1))
feature_names.append("#Rating_2_P" + str(i + 1))
feature_names.append("#Rating_3_P" + str(i + 1))
feature_names.append("#Rating_4_P" + str(i + 1))
feature_names.append("#Rating_5_P" + str(i + 1))
for i in range(num_time_periods):
feature_names.append("#Helpful_1_P" + str(i + 1))
feature_names.append("#Helpful_2_P" + str(i + 1))
feature_names.append("#Helpful_3_P" + str(i + 1))
feature_names.append("#Helpful_4_P" + str(i + 1))
feature_names.append("#Helpful_5_P" + str(i + 1))
for i in range(num_time_periods):
feature_names.append("#Non_Helpful_1_P" + str(i + 1))
feature_names.append("#Non_Helpful_2_P" + str(i + 1))
feature_names.append("#Non_Helpful_3_P" + str(i + 1))
feature_names.append("#Non_Helpful_4_P" + str(i + 1))
feature_names.append("#Non_Helpful_5_P" + str(i + 1))
for i in range(num_time_periods):
feature_names.append("#Sent_+ve_1_P" + str(i + 1))
feature_names.append("#Sent_-ve_1_P" + str(i + 1))
feature_names.append("#Sent_+ve_2_P" + str(i + 1))
feature_names.append("#Sent_-ve_2_P" + str(i + 1))
feature_names.append("#Sent_+ve_3_P" + str(i + 1))
feature_names.append("#Sent_-ve_3_P" + str(i + 1))
feature_names.append("#Sent_+ve_4_P" + str(i + 1))
feature_names.append("#Sent_-ve_4_P" + str(i + 1))
feature_names.append("#Sent_+ve_5_P" + str(i + 1))
feature_names.append("#Sent_-ve_5_P" + str(i + 1))
#for i in range(X.shape[1]):
# feature_names.append("feature_" + str(i + 1))
return feature_names
def Feature_Selection_Using_Feature_Importance(X,Y,category_name,feature_names,num_max_features):
# feature extraction
print("Feature Importance Feature selection")
model = ExtraTreesClassifier(class_weight='balanced',n_jobs=1)
model.fit(X, Y)
#print(model.feature_importances_)
#print(type(model.feature_importances_))
feat_imp = model.feature_importances_
indices = np.argsort(feat_imp)[::-1]
#print(indices)
print("len feature names "+str(len(feature_names)))
selected_feature_names =[]
for f in range(X.shape[1]):
print("%d. feature %d (%f) %s" % (f + 1, indices[f], feat_imp[indices[f]],feature_names[indices[f]]))
if f <num_max_features:
selected_feature_names.append(feature_names[indices[f]])
#print(feature_names)
print (selected_feature_names)
plot_feature_importance2(X, feat_imp, indices,selected_feature_names,category_name,num_max_features) #feature_names,category_name,num_max_features)
return
file_path = "D:\Yassien_PhD\Experiment_6\Train_Test_Category_TQ_Target_Sub_Cat_Setup/train.txt"
features_vectors,ranks = Extract_Features_From_File(file_path)
X = np.array(features_vectors)
Y = np.array(ranks)
print(X.shape)
print(Y.shape)
timeperiods=1
num_max_features = 11
#feature_names = get_feature_names(timeperiods)
feature_names = []
feature_names.append("1-Star")
feature_names.append("2-Star")
feature_names.append("3-Star")
feature_names.append("4-Star")
feature_names.append("5-Star")
feature_names.append("Avg_Star")
feature_names.append("Min LS")
feature_names.append("Max LS")
feature_names.append("Median LS")
feature_names.append("Median # Revs")
feature_names.append("Median Active")
Feature_Selection_Using_Feature_Importance(X,Y,"Jewelry",feature_names,num_max_features)
#Feature_Selection_Using_UniVariate(X,Y)
#Feature_Selection_Using_Recursive_Feature_Elimination(X,Y)
#Feature_Selection_Using_Principal_Component_Analysis(X,Y)