-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtest_dynamic_edges_dense_GAD_find_wrong_subjects.py
269 lines (252 loc) · 14.2 KB
/
test_dynamic_edges_dense_GAD_find_wrong_subjects.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
""" Training and testing of the model
"""
import os
import numpy as np
from sklearn.metrics import accuracy_score, f1_score, roc_auc_score,confusion_matrix,roc_curve
import torch
import torch.nn.functional as F
from models_dynamic_edges_dense_k_hup_neighbors import init_model_dict, init_optim
from utils import one_hot_tensor, cal_sample_weight, gen_adj_mat_tensor, gen_test_adj_mat_tensor, cal_adj_mat_parameter,load_model_dict,save_Intermediate_results
import pandas as pd
import csv
import xlrd
from sklearn.model_selection import StratifiedShuffleSplit
from sklearn.model_selection import StratifiedKFold
from sklearn import preprocessing
cuda = True if torch.cuda.is_available() else False
def prepare_trte_data(data_folder, view_list,fMRI_type,data_subfolder,data_num,atlas,fMRI,sMRI,DIT):
if data_subfolder=='GAD':
label_file="GAD_label.csv"
if data_subfolder=="COBRE":
label_file = "COBRE_label.csv"
source_folder=os.path.join(data_folder,data_subfolder,data_num)
label_csv=pd.read_csv(os.path.join(source_folder,label_file))
if fMRI_type == 1:
data_fMRI_csv = pd.read_csv(
os.path.join(source_folder,"fMRI",atlas,"ReHo_VMHC_fALFF_ALFF","VMHC","VMHC_features.csv"))
elif fMRI_type == 2:
data_fMRI_csv = pd.read_csv(
os.path.join(source_folder, "fMRI", atlas, "ReHo_VMHC_fALFF_ALFF", "ReHo", "ReHo_features.csv"))
elif fMRI_type == 3:
data_fMRI_csv = pd.read_csv(
os.path.join(source_folder, "fMRI", atlas, "ReHo_VMHC_fALFF_ALFF", "ALFF", "ALFF_features.csv"))
data_fMRI_csv_test = pd.read_csv(fMRI)
elif fMRI_type == 4:
data_fMRI_csv = pd.read_csv(
os.path.join(source_folder, "fMRI", atlas, "ReHo_VMHC_fALFF_ALFF", "fALFF", "fALFF_features.csv"))
elif fMRI_type == 5:
data_fMRI_csv = pd.read_csv(
os.path.join(source_folder, "fMRI", atlas, "fc_fisher", "feature_fc.csv"))
data_sMRI_csv = pd.read_csv(
os.path.join(source_folder, "sMRI", atlas, "GMV", "GMV_features.csv"))
data_sMRI_csv_test = pd.read_csv(sMRI)
# data_sMRI_csv = pd.read_csv(
# os.path.join(source_folder, "fMRI", "aal_116","fc_fisher", "feature_fc.csv"))
test_DIT = []
data_DTI=[]
if data_subfolder == "GAD":
data_DTI_csv = pd.read_csv(
os.path.join(source_folder, "DTI", "AllAtlasResults", "WMlabelResults_FA.csv"))
data_DTI = data_DTI_csv.values[:, 1:]
test_DIT = pd.read_csv(DIT).values[:, 1:]
label = np.array(label_csv.values[:, 1:]) # 所有数据的label
data_fMRI = data_fMRI_csv.values[:, 1:]
data_sMRI = data_sMRI_csv.values[:, 1:]
data_sMRI_test = data_sMRI_csv_test.values[:, 1:]
data_fMRI_test = data_fMRI_csv_test.values[:, 1:]
ids = data_fMRI_csv.values[:, 0:1]
ids = np.reshape(ids, (-1,))
ids = ids.tolist()
ids_test = data_sMRI_csv_test.values[:, 0:1]
ids_test = np.reshape(ids_test, (-1,))
ids_test = ids_test.tolist()
return label,data_fMRI,data_sMRI,data_DTI,data_fMRI_test,data_sMRI_test,test_DIT,ids, ids_test
def gen_trte_adj_mat(data_tr_list, data_trte_list, trte_idx, adj_parameter):
adj_metric = "cosine" # cosine distance
adj_train_list = []
adj_test_list = []
for i in range(len(data_tr_list)):
adj_parameter_adaptive = cal_adj_mat_parameter(adj_parameter[i], data_tr_list[i],
adj_metric) # 计算到node_num*edges_per_node的数字大小，
# 接下来所有大于等于这个数字的余弦距离都留下
adj_train_list.append(gen_adj_mat_tensor(data_tr_list[i], adj_parameter_adaptive, adj_metric))
# adj_train_list中全是非0元素了
adj_test_list.append(gen_test_adj_mat_tensor(data_trte_list[i], trte_idx, adj_parameter_adaptive, adj_metric))
return adj_train_list, adj_test_list
def gen_tr_adj_mat(data_tr_list, adj_parameter):
adj_metric = "cosine" # cosine distance
adj_train_list = []
for i in range(len(data_tr_list)):
adj_parameter_adaptive = cal_adj_mat_parameter(adj_parameter, data_tr_list[i],
adj_metric) # 计算到node_num*edges_per_node的数字大小，
# 接下来所有大于等于这个数字的余弦距离都留下
adj_train_list.append(gen_adj_mat_tensor(data_tr_list[i], adj_parameter_adaptive, adj_metric))
# adj_train_list中全是非0元素了
return adj_train_list
def gen_tr_adj_mat_make_new_graph(data_tr_list, adj_parameter): #每一层GCN后更新adj_list
adj_metric = "cosine" # cosine distance
adj_train_list = []
adj_parameter_adaptive = cal_adj_mat_parameter(adj_parameter, data_tr_list,
adj_metric) # 计算到node_num*edges_per_node的数字大小，
# 接下来所有大于等于这个数字的余弦距离都留下
adj_train_list= gen_adj_mat_tensor(data_tr_list, adj_parameter_adaptive, adj_metric)
# adj_train_list中全是非0元素了
return adj_train_list
def test_epoch(data_list, adj_list, te_idx, model_dict, adj_parameter):
for m in model_dict:
model_dict[m].eval()
num_view = len(data_list)
ci_list = []
adj_list_next = []
for i in range(num_view):
ci_list.append(model_dict["C{:}".format(i + 1)](
model_dict["E{:}".format(i + 1)](adj_parameter[i], data_list[i], adj_list[i])))
if num_view >= 2:
c = model_dict["C"](ci_list)
else:
c = ci_list[0]
c = c[te_idx, :]
prob = F.softmax(c, dim=1).data.cpu().numpy()
for i in range(num_view):
adj_list_next.append(gen_tr_adj_mat_make_new_graph(ci_list[i], adj_parameter[i]))
return prob,adj_list_next
def train_test(data_folder, view_list, num_class,
lr_e_pretrain, lr_e, lr_c,
num_epoch_pretrain, num_epoch, fold,
fMRI_type, fold_repeat_all, data_subfolder,
N_SEED_all, adj_parameter, dim_he_list,data_num,atlas,model_folder,type_folder,preprocessing_label,fMRI,sMRI,DIT, save_results="/home/lining/xmn1/results/"):
base_path = ""
if data_subfolder == "GAD":
print("Use fMRI,sMRI,DTI")
elif data_subfolder == 'COBRE':
print("Use fMRI,sMRI")
test_inverval = 1
models_dir = ""
num_view = len(view_list)
dim_hvcdn = pow(num_class, num_view) # 这是干啥的 流程图里那个立方体
source_folder = os.path.join(data_folder, data_subfolder, data_num)
type_folder=type_folder
models_dir = os.path.join(source_folder, "models_"+atlas,model_folder,type_folder)
print("model:",model_folder)
label, data_fMRI, data_sMRI, data_DTI,test_fMRI,test_sMRI,test_DIT,names,test_names = prepare_trte_data(data_folder, view_list, fMRI_type, data_subfolder,data_num,atlas,fMRI=fMRI,sMRI=sMRI,DIT=DIT)
data_fMRI = np.append(data_fMRI,test_fMRI,axis=0)
data_sMRI = np.append(data_sMRI,test_sMRI,axis=0)
for name in test_names:
names.append(name)
if data_subfolder == 'GAD':
data_DTI = np.append(data_DTI,test_DIT,axis=0)
acc_list, sen_list, spe_list, f1_score_list, auc_list = [], [], [], [], []
yuce_0, yu_ce_1 = [], []
label_int = []
if data_subfolder == 'GAD':
label = np.append(label, [[1]], axis=0)
else:
label = np.append(label, [[0]],axis=0)
for i in range(len(label)):
label_int.append((int)(label[i]))
# fold_repeat = fold_repeat_all
# for fold_repeat in range(fold_repeat_all):
for fold_repeat_time, N_SEED in enumerate(N_SEED_all):
fold_repeat=N_SEED
# N_SEED_next = N_SEED_next +10*fold_repeat
skf = StratifiedKFold(n_splits=fold, shuffle=True, random_state=N_SEED)
# skf = StratifiedKFold(n_splits=fold, shuffle=True)
split_time = -1
for train_index, test_index in skf.split(label[0:len(label_int) - 1], label_int[0:len(label_int) - 1]):
data_tr_list = []
data_te_list = []
test_index = np.append(test_index, len(label_int) - 1)
split_time = split_time + 1
labels_tr = label[train_index, :]
labels_te = label[test_index, :]
if data_subfolder == "GAD" and preprocessing_label==1:
data_tr_list.append(preprocessing.scale(data_fMRI[train_index, :]))
data_te_list.append(preprocessing.scale(data_fMRI[test_index, :]))
data_tr_list.append(preprocessing.scale(data_sMRI[train_index, :]))
data_te_list.append(preprocessing.scale(data_sMRI[test_index, :]))
data_tr_list.append(preprocessing.scale(data_DTI[train_index, :]))
data_te_list.append(preprocessing.scale(data_DTI[test_index, :]))
elif data_subfolder == "GAD" and preprocessing_label == 0:
data_tr_list.append(data_sMRI[train_index, :])
data_te_list.append(data_sMRI[test_index, :])
data_tr_list.append(data_DTI[train_index, :])
data_te_list.append(data_DTI[test_index, :])
elif data_subfolder == "COBRE" and preprocessing_label==1:
data_tr_list.append(preprocessing.scale(data_fMRI[train_index, :]))
data_te_list.append(preprocessing.scale(data_fMRI[test_index, :]))
data_tr_list.append(preprocessing.scale(data_sMRI[train_index, :]))
data_te_list.append(preprocessing.scale(data_sMRI[test_index, :]))
elif data_subfolder == "COBRE" and preprocessing_label == 0:
data_tr_list.append(data_fMRI[train_index, :])
data_te_list.append(data_fMRI[test_index, :])
data_tr_list.append(data_sMRI[train_index, :])
data_te_list.append(data_sMRI[test_index, :])
num_view = len(view_list)
labels_tr = labels_tr.astype(int)
labels_te = labels_te.astype(int)
num_tr = data_tr_list[0].shape[0]
num_te = data_te_list[0].shape[0]
data_mat_list = []
for i in range(num_view):
data_mat_list.append(np.concatenate((data_tr_list[i], data_te_list[i]), axis=0))
data_tensor_list = []
for i in range(len(data_mat_list)):
data_tensor_list.append(torch.FloatTensor(data_mat_list[i].astype(float)))
if cuda:
data_tensor_list[i] = data_tensor_list[i].cuda()
idx_dict = {}
idx_dict["tr"] = list(range(num_tr))
idx_dict["te"] = list(range(num_tr, (num_tr + num_te)))
data_train_list = []
data_all_list = []
for i in range(len(data_tensor_list)):
data_train_list.append(data_tensor_list[i][idx_dict["tr"]].clone())
data_all_list.append(torch.cat((data_tensor_list[i][idx_dict["tr"]].clone(),
data_tensor_list[i][idx_dict["te"]].clone()), 0))
labels_trte = np.concatenate((labels_tr, labels_te))
data_tr_list, data_trte_list, trte_idx, labels_trte = data_train_list, data_all_list, idx_dict, labels_trte
labels_tr_tensor = torch.LongTensor(labels_trte[trte_idx["tr"]])
onehot_labels_tr_tensor = one_hot_tensor(labels_tr_tensor, num_class)
sample_weight_tr = cal_sample_weight(labels_trte[trte_idx["tr"]], num_class)
# 为解决样本标签不平衡问题，sample_weight_tr为每一个训练样本设置一个权重，应用于每个模态GCN的损失函数中不同类别的损失，权重为其在训练数据中频率的倒数
sample_weight_tr = torch.FloatTensor(sample_weight_tr)
if cuda:
labels_tr_tensor = labels_tr_tensor.cuda()
onehot_labels_tr_tensor = onehot_labels_tr_tensor.cuda()
sample_weight_tr = sample_weight_tr.cuda()
adj_tr_list, adj_te_list = gen_trte_adj_mat(data_tr_list, data_trte_list, trte_idx, adj_parameter)
dim_list = [x.shape[1] for x in data_tr_list]
# 每一模态的初始特征维度[116, 116, 48]
model_dict = init_model_dict(num_view, num_class, dim_list, dim_he_list, dim_hvcdn)
for m in model_dict:
if cuda:
model_dict[m].cuda()
model_dict=load_model_dict(models_dir, model_dict,fold_repeat,split_time)
te_prob,adj_test = test_epoch(data_trte_list, adj_te_list, trte_idx["te"], model_dict, adj_parameter)
_ = save_Intermediate_results(v=names,adj_list=adj_te_list,num=split_time,type="original",type_shujv=data_subfolder,base_path=save_results,name=names[-1])
base_path = save_Intermediate_results(v=names,adj_list=adj_test,num=split_time,type="Intermediate",type_shujv=data_subfolder,base_path=save_results,name=names[-1])
if num_class == 2:
a = labels_trte[trte_idx["te"]]
tn, fp, fn, tp = confusion_matrix(labels_trte[trte_idx["te"]], te_prob.argmax(1)).ravel()
acc = accuracy_score(labels_trte[trte_idx["te"]], te_prob.argmax(1))
sen = tp / (tp + fn)
spe = tn / (tn + fp)
auc = roc_auc_score(labels_trte[trte_idx["te"]], te_prob[:, 1])
f1_scor = f1_score(labels_trte[trte_idx["te"]], te_prob.argmax(1))
fpr, tpr, threshold = roc_curve(labels_trte[trte_idx["te"]], te_prob[:, 1])
yuce_0.append(te_prob[-1][0])
yu_ce_1.append(te_prob[-1][1])
acc_list.append(acc)
sen_list.append(sen)
spe_list.append(spe)
f1_score_list.append(f1_scor)
auc_list.append(auc)
predicts = os.path.join(save_results,data_subfolder,names[-1],"predict.txt")
with open(predicts, 'w') as f:
f.write('Probability of depression:{:.4f}'.format(np.mean(yu_ce_1)))
f.write("\n")
if np.mean(yu_ce_1) < 0.5:
f.write('Subject health\n')
else:
f.write('Subject suffers from depression\n')
#print("Intermediate results are stored in:",base_path)