|
@@ -6,20 +6,21 @@ from keras.layers import Dense,Dropout
|
6
|
6
|
import random
|
7
|
7
|
from keras.models import load_model
|
8
|
8
|
|
9
|
|
-lines = []
|
|
9
|
+
|
10
|
10
|
def read_data(path):
|
|
11
|
+ lines = []
|
11
|
12
|
with open(path) as f:
|
12
|
|
- for line in f.readlines():
|
|
13
|
+ for line in f.readlines()[:]:
|
13
|
14
|
lines.append(eval(line.strip()))
|
14
|
15
|
|
15
|
16
|
size = len(lines[0])
|
16
|
17
|
train_x=[s[:size - 2] for s in lines]
|
17
|
18
|
train_y=[s[size-1] for s in lines]
|
18
|
|
- return np.array(train_x),np.array(train_y)
|
|
19
|
+ return np.array(train_x),np.array(train_y),lines
|
19
|
20
|
|
20
|
21
|
|
21
|
22
|
def predict(file_path='', model_path='15min_dnn_seq.h5'):
|
22
|
|
- test_x,test_y=read_data(file_path)
|
|
23
|
+ test_x,test_y,lines=read_data(file_path)
|
23
|
24
|
|
24
|
25
|
model=load_model(model_path)
|
25
|
26
|
score = model.evaluate(test_x, test_y)
|
|
@@ -35,6 +36,7 @@ def predict(file_path='', model_path='15min_dnn_seq.h5'):
|
35
|
36
|
fact = test_y[i]
|
36
|
37
|
if r[0] > 0.5:
|
37
|
38
|
f.write(str([lines[i][-2], lines[i][-1]]) + "\n")
|
|
39
|
+ win_dnn.append([lines[i][-2], lines[i][-1]])
|
38
|
40
|
if fact[0] == 1:
|
39
|
41
|
up_right = up_right + 1
|
40
|
42
|
elif fact[1] == 1:
|
|
@@ -47,5 +49,5 @@ def predict(file_path='', model_path='15min_dnn_seq.h5'):
|
47
|
49
|
|
48
|
50
|
|
49
|
51
|
if __name__ == '__main__':
|
50
|
|
- # predict(file_path='D:\\data\\quantization\\stock6_5_test.log', model_path='5d_dnn_seq.h5')
|
51
|
|
- predict(file_path='D:\\data\\quantization\\stock6_test.log', model_path='15m_dnn_seq.h5')
|
|
52
|
+ predict(file_path='D:\\data\\quantization\\stock6_5_test.log', model_path='5d_dnn_seq.h5')
|
|
53
|
+ # predict(file_path='D:\\data\\quantization\\stock6_test.log', model_path='15m_dnn_seq.h5')
|