-
Notifications
You must be signed in to change notification settings - Fork 0
/
feature_selection_sp.py
126 lines (112 loc) · 5.02 KB
/
feature_selection_sp.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
import re
import os
import numpy as np
from sklearn.feature_selection import chi2
def scan_token(infile,ofile,d):
o=open(ofile,'w+')
f=open(infile,'r')
line=f.readline().strip()
o.write(line+'\n')
while True:
line=f.readline().strip()
if not line:break
ele=line.split('\t')
tk=re.split(',',ele[-1])
if len(tk)==1 and tk[0]=='0':
o.write(line+'\n')
continue
tem=[]
for t in tk:
if t=='0':continue
if t in d:
tem.append(t)
if len(tem)==0:
o.write(ele[0]+'\t'+ele[1]+'\t'+str(len(tem))+'\t0\n')
else:
o.write(ele[0]+'\t'+ele[1]+'\t'+str(len(tem))+'\t'+','.join(tem)+'\n')
def sef(infile,ofile,ofile2):
f=open(infile,'r')
line=f.readline()
d={} # strain -> token_id -> frequency
arr=[] # token_id list
s=[] # strain_id list
y=[]
while True:
line=f.readline().strip()
if not line:break
ele=line.split('\t')
d[ele[0]]={}
y.append(int(ele[1]))
s.append(ele[0])
tk=re.split(',',ele[-1])
for e in tk:
if e not in d[ele[0]]:
d[ele[0]][e]=1
else:
d[ele[0]][e]+=1
if e not in arr:
arr.append(e)
X=[]
for x in s:
tem=[]
for a in arr:
if a not in d[x]:
tem.append(0)
else:
tem.append(d[x][a])
X.append(tem)
X=np.array(X)
y=np.array(y)
#sel = VarianceThreshold(threshold=(.8 * (1 - .8)))
#nX=sel.fit_transform(X)
#print(nX.shape)
scores, pvalues = chi2(X, y)
o=open(ofile,'w+')
o.write('ID\tFeature_ID\tP-value\tChi2-statistic\n')
c=1
dr={}
di={}
all_t={}
for i in range(len(scores)):
if pvalues[i]>0.05:
all_t[arr[i]]=''
continue
#print(f"Feature {arr[i]}: P-value = {pvalues[i]}, Chi2-statistic = {scores[i]}")
dr[arr[i]]=pvalues[i]
#o.write(str(c)+'\t'+str(arr[i])+'\t'+str(pvalues[i])+'\t'+str(scores[i])+'\n')
di[arr[i]]=str(arr[i])+'\t'+str(pvalues[i])+'\t'+str(scores[i])+'\n'
#c+=1
#exit()
res=sorted(dr.items(), key = lambda kv:(kv[1], kv[0]))
dused={}
for r in res:
o.write(str(c)+'\t'+di[r[0]])
dused[r[0]]=''
c+=1
if len(res)==0:
dused=all_t
scan_token(infile,ofile2,dused)
'''
o2=open(ofile2,'w+')
f2=open(infile,'r')
line=f2.readline()
o2.write(line)
while True:
line=f2.readline().strip()
if not line:break
ele=line.split('\t')
tk=re.split(',',ele[-1])
tem=[]
for t in tk:
if t in dused:
tem.append(t)
#ele[-2]=str(len(tem))
o2.write(ele[0]+'\t'+ele[1]+'\t'+ele[2]+'\t'+str(len(tem))+'\t'+','.join(tem)+'\n')
'''
#scan_token(infile2,ov,dused)
#sef('Build_multimodal_tokens/Kcp_3fold/Fold1/strains_train_sentence.txt','Build_multimodal_tokens/Kcp_3fold/Fold1/strains_val_sentence.txt','Build_multimodal_tokens/Kcp_3fold/Fold1/kcp_feature_remain_graph.txt','Build_multimodal_tokens/Kcp_3fold/Fold1/strains_train_sentence_fs.txt','Build_multimodal_tokens/Kcp_3fold/Fold1/strains_val_sentence_fs.txt')
#sef('Build_multimodal_tokens/Kcp_3fold/Fold1/strains_train_pc_token.txt','Build_multimodal_tokens/Kcp_3fold/Fold1/strains_val_pc_token.txt','Build_multimodal_tokens/Kcp_3fold/Fold1/kcp_feature_remain_pc.txt','Build_multimodal_tokens/Kcp_3fold/Fold1/strains_train_pc_token_fs.txt','Build_multimodal_tokens/Kcp_3fold/Fold1/strains_val_pc_token_fs.txt')
#sef('Build_multimodal_tokens/Kcp_3fold/Fold2/strains_train_sentence.txt','Build_multimodal_tokens/Kcp_3fold/Fold2/strains_val_sentence.txt','Build_multimodal_tokens/Kcp_3fold/Fold2/kcp_feature_remain_graph.txt','Build_multimodal_tokens/Kcp_3fold/Fold2/strains_train_sentence_fs.txt','Build_multimodal_tokens/Kcp_3fold/Fold2/strains_val_sentence_fs.txt')
#sef('Build_multimodal_tokens/Kcp_3fold/Fold2/strains_train_pc_token.txt','Build_multimodal_tokens/Kcp_3fold/Fold2/strains_val_pc_token.txt','Build_multimodal_tokens/Kcp_3fold/Fold2/kcp_feature_remain_pc.txt','Build_multimodal_tokens/Kcp_3fold/Fold2/strains_train_pc_token_fs.txt','Build_multimodal_tokens/Kcp_3fold/Fold2/strains_val_pc_token_fs.txt')
#sef('Build_multimodal_tokens/Kcp_3fold/Fold3/strains_train_sentence.txt','Build_multimodal_tokens/Kcp_3fold/Fold3/strains_val_sentence.txt','Build_multimodal_tokens/Kcp_3fold/Fold3/kcp_feature_remain_graph.txt','Build_multimodal_tokens/Kcp_3fold/Fold3/strains_train_sentence_fs.txt','Build_multimodal_tokens/Kcp_3fold/Fold3/strains_val_sentence_fs.txt')
#sef('Build_multimodal_tokens/Kcp_3fold/Fold3/strains_train_pc_token.txt','Build_multimodal_tokens/Kcp_3fold/Fold3/strains_val_pc_token.txt','Build_multimodal_tokens/Kcp_3fold/Fold3/kcp_feature_remain_pc.txt','Build_multimodal_tokens/Kcp_3fold/Fold3/strains_train_pc_token_fs.txt','Build_multimodal_tokens/Kcp_3fold/Fold3/strains_val_pc_token_fs.txt')