-
Notifications
You must be signed in to change notification settings - Fork 6
/
Copy pathutils.py
174 lines (148 loc) · 5.73 KB
/
utils.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
#encoding:utf-8
import numpy as np
from scipy.io import loadmat
from scipy.sparse import csc_matrix
from sklearn.preprocessing import normalize
from sklearn.metrics import f1_score
import scipy.sparse as sp
from collections import defaultdict
from sklearn.metrics import accuracy_score,roc_auc_score,average_precision_score,precision_score,jaccard_similarity_score,recall_score,zero_one_loss
import torch
def make_attribute():
path1 = '../../dataHIN/DBLP2/PA2.txt'
path2 = '../../dataHIN/DBLP2/PT2.txt'
paper_term_map={}
term_set=set()
with open(path2) as fp:
for line in fp.readlines():
paper=line.strip('\n\r').split()[0]
term=line.strip('\n\r').split()[1]
if paper not in paper_term_map:
paper_term_map[paper]=[]
paper_term_map[paper].append(term)
term_set.add(term)
author_term_map={}
with open(path1) as fp:
for line in fp.readlines():
author=line.strip('\n\r').split()[1]
paper=line.strip('\n\r').split()[0]
term_list=paper_term_map[paper]
if author not in author_term_map:
author_term_map[author]=[]
author_term_map[author].append(term_list)
print('Term number:',len(term_set))
print('Author number:',len(author_term_map))
feature_mat=np.zeros((len(author_term_map),len(term_set)))
for author in author_term_map:
term_lists=author_term_map[author]
for term_list in term_lists:
for term in term_list:
feature_mat[int(author),int(term)]=1
res=normalize(feature_mat,norm='l2')
return res
def hamming_score(y_true, y_pred, normalize=True, sample_weight=None):
'''
Compute the Hamming score (a.k.a. label-based accuracy) for the multi-label case
https://stackoverflow.com/q/32239577/395857
'''
acc_list = []
for i in range(y_true.shape[0]):
set_true = set( np.where(y_true[i])[0] )
set_pred = set( np.where(y_pred[i])[0] )
#print('\nset_true: {0}'.format(set_true))
#print('set_pred: {0}'.format(set_pred))
tmp_a = None
if len(set_true) == 0 and len(set_pred) == 0:
tmp_a = 1
else:
tmp_a = len(set_true.intersection(set_pred))/\
float( len(set_true.union(set_pred)) )
#print('tmp_a: {0}'.format(tmp_a))
acc_list.append(tmp_a)
return np.mean(acc_list)
def multilabel_f1(GT,pred):
labeled_num_list=torch.sum(GT,dim=1)
for i in range(len(labeled_num_list)):
if labeled_num_list[i]==0:
labeled_num_list[i]=1
pred_label=torch.zeros_like(GT)
for i in range(pred.shape[0]):
pred_label[i,torch.topk(pred[i],int(labeled_num_list[i]))[1]]=1
#score1 = accuracy_score(y_true=GT,y_pred=pred_label)
#score1 = roc_auc_score(y_true=GT,y_score=pred)
score1=0
score2 = average_precision_score(y_true=GT,y_score=pred,average='micro')
#score = hamming_loss(GT,pred_label)
#score = precision_score(GT,pred_label,average='micro')
#print(GT,pred_label)
# a_list=[]
# b_list=[]
# for i in range(GT.shape[0]):
# aa=f1_score(GT[i, :], pred_label[i, :], average='micro')
# bb=f1_score(GT[i, :], pred_label[i, :], average='macro')
# a_list.append(aa)
# b_list.append(bb)
# return np.mean(a_list),np.mean(b_list)
#return hamming_score(GT,pred_label),hamming_score(GT,pred_label)
return f1_score(GT,pred_label,average='micro'),f1_score(GT,pred_label,average='macro'),score2,score1
def normalize_adj(mx):
"""Row-normalize sparse matrix"""
rowsum = np.array(mx.sum(1))
r_inv_sqrt = np.power(rowsum, -0.5).flatten()
r_inv_sqrt[np.isinf(r_inv_sqrt)] = 0.
r_mat_inv_sqrt = sp.diags(r_inv_sqrt)
return mx.dot(r_mat_inv_sqrt).transpose().dot(r_mat_inv_sqrt)
def adj2Lap(adj):
AA=adj+np.ones_like(adj)
D=np.diag(np.diag(AA))
DD=np.zeros_like(D)
[x,y]=np.nonzero(D)
for i in range(len(x)):
DD[x[i],y[i]]=np.power(D[x[i],y[i]],-0.5)
return np.dot(np.dot(DD,AA),DD)
def get_adj(path):
edge_mat = np.loadtxt(path, dtype=int)
node_num = int(np.max(edge_mat)) + 1
adj_mat = np.zeros((node_num, node_num), dtype=int)
adj_lists = defaultdict(set)
for i in range(edge_mat.shape[0]):
adj_mat[edge_mat[i, 0], edge_mat[i, 1]] = 1
adj_mat[edge_mat[i, 1], edge_mat[i, 0]] = 1
adj_lists[edge_mat[i, 0]].add(edge_mat[i, 1])
adj_lists[edge_mat[i, 1]].add(edge_mat[i, 0])
adj_lists[edge_mat[i, 0]].add(edge_mat[i, 0])
adj_lists[edge_mat[i, 1]].add(edge_mat[i, 1])
for i in range(node_num):
if adj_mat[i,i]==0:
adj_mat[i, i] = 1
return adj_mat, adj_lists
def get_adj_mat(path,name):
adj_mat=loadmat(path)[name].todense()
for i in range(adj_mat.shape[0]):
adj_mat[i,i]=1
adj_lists = defaultdict(set)
[row,col]=adj_mat.nonzero()
for i in range(len(row)):
node1=row[i]
node2=col[i]
adj_lists[node1].add(node2)
adj_lists[node1].add(node1)
adj_lists[node2].add(node1)
adj_lists[node2].add(node2)
return adj_mat,adj_lists
def get_label(path):
label_mat = np.loadtxt(path, dtype=int)
node_num = label_mat.shape[0]
num_class = int(np.max(label_mat[:, 1]) + 1)
label_list = np.array(label_mat[:, 1])
return num_class, node_num, label_list
def get_label2(path):
input_mat = np.loadtxt(path,dtype=int)
node_num = int(np.max(input_mat[:,0])+1)
num_class = int(np.max(input_mat[:,1])+1)
label_mat = np.zeros((node_num,num_class))
for i in range(input_mat.shape[0]):
node = input_mat[i, 0]
label = input_mat[i, 1]
label_mat[node,label]=1
return num_class,node_num,label_mat