forked from scz760904126/AMMGC
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmain_multiGCN_attention.py
326 lines (259 loc) · 12.4 KB
/
main_multiGCN_attention.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
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
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
from __future__ import division
from __future__ import print_function
from datetime import datetime
from models import *
from utils import *
from metrics import *
from attention import attentionModel
import tensorflow.compat.v1 as tf
import random
import matplotlib.pyplot as plt
tf.disable_eager_execution()
def GCN_process(train_drug_miRNA_matrix, adj_train, adj_norm, features, num_nodes, num_edges, positive_train,
positive_mask):
# Set random seed
seed = 123
np.random.seed(seed)
tf.set_random_seed(seed)
# Settings
flags = tf.app.flags
FLAGS = flags.FLAGS
flags.DEFINE_float('learning_rate', 0.01, 'Initial learning rate.')
flags.DEFINE_integer('epochs', 5, 'Number of epochs to train.') # 100 epochs
flags.DEFINE_integer('hidden1', 128, 'Number of units in hidden layer 1.')
flags.DEFINE_integer('hidden2', 128, 'Number of units in hidden layer 2.')
flags.DEFINE_float('dropout', 0.1, 'Dropout rate (1 - keep probability).')
num_drug = 106
num_miRNA = 754
features = normalize_features(features)
features = sparse_to_tuple(sp.coo_matrix(features))
num_features = features[2][1]
features_nonzero = features[1].shape[0]
# Define placeholders
placeholders = {
'features': tf.sparse_placeholder(tf.float32),
'adj_norm': tf.sparse_placeholder(tf.float32),
'adj_label': tf.sparse_placeholder(tf.float32),
'dropout': tf.placeholder_with_default(0., shape=()),
'positive_mask': tf.placeholder(shape=[79924, 1], dtype=tf.int32),
'negative_mask': tf.placeholder(shape=[79924, 1], dtype=tf.int32),
}
# Create model
model = GCNModel(placeholders, num_features, features_nonzero, num_nodes, num_edges, name='yeast_gcn')
# Initialize session
sess = tf.Session()
sess.run(tf.global_variables_initializer())
adj_label = adj_train + sp.eye(adj_train.shape[0])
adj_label = adj_label.todense()
adj_label = adj_label[0:106, 106::]
adj_label = sp.csr_matrix(adj_label)
adj_label = sparse_to_tuple(adj_label)
epoches = []
avg_costs = []
# Train model
for epoch in range(FLAGS.epochs):
# Create optimizer
negative_mask = generate_mask(train_drug_miRNA_matrix, len(positive_train))
t = time.time()
# Construct feed dictionary
feed_dict = construct_feed_dict(adj_norm, adj_label, features, positive_mask, negative_mask, placeholders)
feed_dict.update({placeholders['dropout']: FLAGS.dropout})
# One update of parameter matrices
_, avg_cost = sess.run([model.opt_op, model.cost], feed_dict=feed_dict)
epoches.append(epoch)
avg_costs.append(avg_cost)
print("Epoch:", '%04d' % (epoch + 1),
"train_loss=", "{:.5f}".format(avg_cost),
"time=", "{:.5f}".format(time.time() - t))
print('Optimization Finished!')
feed_dict.update({placeholders['dropout']: 0})
emb = sess.run(model.embeddings, feed_dict=feed_dict)
def del_all_flags(FLAGS):
flags_dict = FLAGS._flags()
keys_list = [keys for keys in flags_dict]
for keys in keys_list:
FLAGS.__delattr__(keys)
# delete all of flags before running the main command
del_all_flags(flags.FLAGS)
return emb
def attention_process(emb, positive_train, positive_mask, train_drug_miRNA_matrix, num_edges):
# Set random seed
seed = 123
np.random.seed(seed)
tf.set_random_seed(seed)
# Settings
flags = tf.app.flags
FLAGS = flags.FLAGS
flags.DEFINE_string('model', 'attSemiGAE', 'Model string.') # 'gcn', 'semiencoder', 'attSemiGAE'
flags.DEFINE_float('learning_rate', 0.01, 'Initial learning rate.')
flags.DEFINE_integer('epochs', 5, 'Number of epochs to train.') # 300 epochs
flags.DEFINE_float('dropout', 0.5, 'Dropout rate (1 - keep probability).')
flags.DEFINE_float('weight_decay', 0, 'Weight for L2 loss on embedding matrix.')
flags.DEFINE_integer('early_stopping', 50, 'Tolerance for early stopping (# of epochs).')
flags.DEFINE_integer('max_degree', 3, 'Maximum Chebyshev polynomial degree.')
num_supports = len(emb)
# Load data
placeholders = {
'emb': [tf.placeholder(tf.float32, shape=(860, 128)) for _ in range(num_supports)],
'adj_label': tf.placeholder(tf.float32, shape=(106, 754)),
'positive_mask': tf.placeholder(shape=[79924, 1], dtype=tf.int32),
'negative_mask': tf.placeholder(shape=[79924, 1], dtype=tf.int32),
'dropout': tf.placeholder_with_default(0., shape=()),
}
# Create model
model = attentionModel(placeholders, num_edges, logging=True)
# Initialize session
sess = tf.Session()
# Init variables
sess.run(tf.global_variables_initializer())
epoches = []
avg_costs = []
# Train model
for epoch in range(FLAGS.epochs):
negative_mask = generate_mask(train_drug_miRNA_matrix, len(positive_train))
t = time.time()
# Construct feed dictionary
feed_dict = construct_attention_feed_dict(emb, train_drug_miRNA_matrix, positive_mask, negative_mask,
placeholders)
feed_dict.update({placeholders['dropout']: FLAGS.dropout})
# Training step
outs = sess.run([model.opt_op, model.loss], feed_dict=feed_dict)
epoches.append(epoch)
avg_costs.append(outs[1])
# Print results
print("Epoch:", '%04d' % (epoch + 1), "train_loss=", "{:.5f}".format(outs[1]), "time=",
"{:.5f}".format(time.time() - t))
# name = 'gcn_result_con/5_fold_epoches.csv'
# np.savetxt(name, epoches, delimiter=',')
# name = 'gcn_result_con/5_fold_avg_costs.csv'
# np.savetxt(name, avg_costs, delimiter=',')
print("Optimization Finished!")
feed_dict.update({placeholders['dropout']: 0})
output = sess.run(model.output, feed_dict=feed_dict)
def del_all_flags(FLAGS):
flags_dict = FLAGS._flags()
keys_list = [keys for keys in flags_dict]
for keys in keys_list:
FLAGS.__delattr__(keys)
del_all_flags(flags.FLAGS)
return output
# 交叉验证
def cross_validation(adj, seed):
num_nodes = adj.shape[0]
drug_feature_name1 = 'features/drug_feature_matrix.txt'
drug_feature1 = np.loadtxt(drug_feature_name1, dtype=float)
drug_feature_name2 = 'features/drug_labelencoding.txt'
drug_feature2 = np.loadtxt(drug_feature_name2, dtype=float)
miRNA_feature_name1 = 'features/ncrna_expression_full.txt'
miRNA_feature1 = np.loadtxt(miRNA_feature_name1, dtype=float)
miRNA_feature_name2 = 'features/ncrna_GOsimilarity_full.txt'
miRNA_feature2 = np.loadtxt(miRNA_feature_name2, dtype=float)
features1 = np.vstack((drug_feature1, np.hstack((np.zeros(
shape=(miRNA_feature1.shape[0], drug_feature1.shape[1] - miRNA_feature1.shape[1]), dtype=int),
miRNA_feature1))))
features2 = np.vstack((np.hstack((np.zeros(
shape=(drug_feature1.shape[0], miRNA_feature2.shape[1] - drug_feature1.shape[1]), dtype=int), drug_feature1)),
miRNA_feature2))
features3 = np.vstack((np.hstack((np.zeros(
shape=(drug_feature2.shape[0], miRNA_feature1.shape[1] - drug_feature2.shape[1]), dtype=int), drug_feature2)),
miRNA_feature1))
features4 = np.vstack((np.hstack((np.zeros(
shape=(drug_feature2.shape[0], miRNA_feature2.shape[1] - drug_feature2.shape[1]), dtype=int), drug_feature2)),
miRNA_feature2))
num_drug = 106
drug_miRNA_matrix = adj.todense()[0:num_drug, num_drug::]
none_zero_position = np.where(drug_miRNA_matrix != 0)
none_zero_row_index = none_zero_position[0]
none_zero_col_index = none_zero_position[1]
np.random.seed(seed)
positive_randomlist = [i for i in range(len(none_zero_row_index))]
random.shuffle(positive_randomlist)
sum_metric = np.zeros((1, 7))
k_folds = 5
print("seed=%d, evaluating miRNA-disease...." % (seed))
for k in range(k_folds):
metric = np.zeros((1, 7))
print("------this is %dth cross validation------" % (k + 1))
if k != k_folds - 1:
positive_test = positive_randomlist[k * int(len(none_zero_row_index) / k_folds):(k + 1) * int(
len(none_zero_row_index) / k_folds)]
positive_train = list(set(positive_randomlist).difference(set(positive_test)))
else:
positive_test = positive_randomlist[k * int(len(none_zero_row_index) / k_folds)::]
positive_train = list(set(positive_randomlist).difference(set(positive_test)))
positive_test_row = none_zero_row_index[positive_test]
positive_test_col = none_zero_col_index[positive_test]
train_drug_miRNA_matrix = np.copy(drug_miRNA_matrix)
train_drug_miRNA_matrix[positive_test_row, positive_test_col] = 0
positive_mask = train_drug_miRNA_matrix.reshape(-1, 1)
num_edges = train_drug_miRNA_matrix.sum()
print('训练集中边的数目')
print(num_edges)
adj_train = constructNet(train_drug_miRNA_matrix)
adj_train = sp.csr_matrix(adj_train)
# 得到训练矩阵
adj = adj_train
adj_norm = preprocess_graph(adj)
tf.reset_default_graph()
emb1 = GCN_process(train_drug_miRNA_matrix, adj_train, adj_norm, features1, num_nodes, num_edges,
positive_train, positive_mask)
emb2 = GCN_process(train_drug_miRNA_matrix, adj_train, adj_norm, features2, num_nodes, num_edges,
positive_train, positive_mask)
emb3 = GCN_process(train_drug_miRNA_matrix, adj_train, adj_norm, features3, num_nodes, num_edges,
positive_train, positive_mask)
emb4 = GCN_process(train_drug_miRNA_matrix, adj_train, adj_norm, features4, num_nodes, num_edges,
positive_train, positive_mask)
emb = []
emb.append(emb1)
emb.append(emb2)
emb.append(emb3)
emb.append(emb4)
tf.reset_default_graph()
att_emb = attention_process(emb, positive_train, positive_mask, train_drug_miRNA_matrix, num_edges)
name = 'gcn_result_con/5_fold_att_emb.csv'
np.savetxt(name, att_emb, delimiter=',')
adj_rec = np.dot(att_emb, att_emb.T)
adj_rec = adj_rec[0: num_drug, num_drug::]
positive_test_row = none_zero_row_index[positive_test]
positive_test_col = none_zero_col_index[positive_test]
negative_position = np.where(drug_miRNA_matrix == 0)
negative_test_row = negative_position[0]
negative_test_col = negative_position[1]
test_row = np.append(positive_test_row, negative_test_row)
test_col = np.append(positive_test_col, negative_test_col)
test_real = []
test_pre = []
for i in range(len(test_row)):
label = drug_miRNA_matrix[test_row[i], test_col[i]]
score = sigmoid(adj_rec[test_row[i], test_col[i]])
test_real.append(label)
test_pre.append(score)
test_real = np.array(test_real)
test_pre = np.array(test_pre)
metric = model_evaluate(test_real, test_pre)
print("------the metrics of %dth cross validation------" % (k + 1))
print(metric)
sum_metric += metric
return sum_metric / k_folds
datetime1 = datetime.now()
name = 'drug-miRNA.txt'
bi_adj = np.zeros((106, 754))
index = np.loadtxt(name, dtype=int)
row = index[:, 0] - 1
col = index[:, 1] - 1
bi_adj[row, col] = 1
adj_dense = constructNet(bi_adj)
adj = sp.csr_matrix(adj_dense)
sum_metric = np.zeros((1, 7))
circle = 10
for i in range(circle):
metric = np.zeros((1, 7))
metric = cross_validation(adj, i)
sum_metric += metric
name = 'gcn_result_con/128_128_multi_attention_5fold_drug_miRNA_seed' + str(i) + '.csv'
np.savetxt(name, metric, delimiter=',')
avg_metric = sum_metric / circle
name = 'gcn_result_con/avg_128_128_multi_attention_5fold_drug_miRNA.csv'
np.savetxt(name, avg_metric, delimiter=',')
print('##########running time###############')
print(datetime.now() - datetime1)