|
@@ -0,0 +1,52 @@
|
|
1
|
+import keras
|
|
2
|
+# -*- encoding:utf-8 -*-
|
|
3
|
+import numpy as np
|
|
4
|
+from keras.models import Sequential
|
|
5
|
+from keras.layers import Dense,Dropout
|
|
6
|
+import random
|
|
7
|
+from keras.models import load_model
|
|
8
|
+
|
|
9
|
+def read_data(path):
|
|
10
|
+ train_x=[]
|
|
11
|
+ train_y=[]
|
|
12
|
+ lines = []
|
|
13
|
+ with open(path) as f:
|
|
14
|
+ for x in range(150000):
|
|
15
|
+ lines.append(eval(f.readline().strip()))
|
|
16
|
+ random.shuffle(lines)
|
|
17
|
+ lines = lines[:20000]
|
|
18
|
+ d=int(0.95*len(lines))
|
|
19
|
+
|
|
20
|
+ size = len(lines[0])
|
|
21
|
+ train_x=[s[:size - 2] for s in lines[0:d]]
|
|
22
|
+ train_y=[s[size-1] for s in lines[0:d]]
|
|
23
|
+ test_x=[s[:size - 2] for s in lines[d:]]
|
|
24
|
+ test_y=[s[size-1] for s in lines[d:]]
|
|
25
|
+ return np.array(train_x),np.array(train_y),np.array(test_x),np.array(test_y)
|
|
26
|
+
|
|
27
|
+train_x,train_y,test_x,test_y=read_data("D:\\data\\quantization\\stock5.log")
|
|
28
|
+
|
|
29
|
+model = Sequential()
|
|
30
|
+model.add(Dense(units=325, input_dim=163, activation='relu'))
|
|
31
|
+model.add(Dense(units=225, activation='relu'))
|
|
32
|
+model.add(Dense(units=125, activation='relu'))
|
|
33
|
+# model.add(Dropout(0.2)(Dense(units=225, activation='relu')))
|
|
34
|
+model.add(Dense(units=8, activation='softmax'))
|
|
35
|
+model.compile(loss='categorical_crossentropy', optimizer="adam",metrics=['accuracy'])
|
|
36
|
+
|
|
37
|
+print("Starting training ")
|
|
38
|
+h=model.fit(train_x, train_y, batch_size=16, epochs=10, shuffle=True)
|
|
39
|
+score = model.evaluate(test_x, test_y)
|
|
40
|
+print(score)
|
|
41
|
+print('Test score:', score[0])
|
|
42
|
+print('Test accuracy:', score[1])
|
|
43
|
+
|
|
44
|
+path="model_seq.h5"
|
|
45
|
+model.save(path)
|
|
46
|
+model=None
|
|
47
|
+
|
|
48
|
+
|
|
49
|
+model=load_model(path)
|
|
50
|
+result=model.predict(test_x)
|
|
51
|
+print(result)
|
|
52
|
+print(test_y)
|