-
Notifications
You must be signed in to change notification settings - Fork 6
/
Copy pathgan_zimbrao.py
267 lines (208 loc) · 10.9 KB
/
gan_zimbrao.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
# original reference: https://github.com/Zackory/Keras-MNIST-GAN
import sys
import os
import shutil
from keras.datasets import mnist
import numpy as np
import math
from keras.optimizers import Adam
from keras.initializers import RandomNormal
from keras.models import Model, Sequential
from keras.layers import Reshape, Dense, Dropout, Flatten, Conv2D, LeakyReLU, Activation, Input, concatenate
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import matplotlib.finance
import time
import pandas as pd
from sklearn.metrics import mean_squared_error
np.random.seed(7)
train = pd.read_csv('minidolar/train.csv', sep = ',', engine='python', decimal='.',header=0)
test = pd.read_csv('minidolar/test.csv', sep = ',', engine='python', decimal='.',header=0)
train_shift = train['shift']
train_target = train['f0']
train_open = train[['v0','v4','v8','v12','v16','v20','v24','v28','v32','v36','v40','v44','v48','v52','v56','v60','v64','v68','v72','v76','v80','v84','v88','v92','v96','v100','v104','v108','v112','v116']]
train_high = train[['v1','v5','v9','v13','v17','v21','v25','v29','v33','v37','v41','v45','v49','v53','v57','v61','v65','v69','v73','v77','v81','v85','v89','v93','v97','v101','v105','v109','v113','v117']]
train_low = train[['v2','v6','v10','v14','v18','v22','v26','v30','v34','v38','v42','v46','v50','v54','v58','v62','v66','v70','v74','v78','v82','v86','v90','v94','v98','v102','v106','v110','v114','v118']]
train_close = train[['v3','v7','v11','v15','v19','v23','v27','v31','v35','v39','v43','v47','v51','v55','v59','v63','v67','v71','v75','v79','v83','v87','v91','v95','v99','v103','v107','v111','v115','v119']]
test_shift = test['shift']
test_target = test['f0']
test_open = test[['v0','v4','v8','v12','v16','v20','v24','v28','v32','v36','v40','v44','v48','v52','v56','v60','v64','v68','v72','v76','v80','v84','v88','v92','v96','v100','v104','v108','v112','v116']]
test_high = test[['v1','v5','v9','v13','v17','v21','v25','v29','v33','v37','v41','v45','v49','v53','v57','v61','v65','v69','v73','v77','v81','v85','v89','v93','v97','v101','v105','v109','v113','v117']]
test_low = test[['v2','v6','v10','v14','v18','v22','v26','v30','v34','v38','v42','v46','v50','v54','v58','v62','v66','v70','v74','v78','v82','v86','v90','v94','v98','v102','v106','v110','v114','v118']]
test_close = test[['v3','v7','v11','v15','v19','v23','v27','v31','v35','v39','v43','v47','v51','v55','v59','v63','v67','v71','v75','v79','v83','v87','v91','v95','v99','v103','v107','v111','v115','v119']]
# obtendo train target para O,H,L
train_l_target = []
train_h_target = []
train_o_target = []
for i in range(train_shift.shape[0]-1):
train_l_target.append(np.array(train_low)[i+1][29]+train_shift[i+1]-train_shift[i])
train_h_target.append(np.array(train_high)[i+1][29]+train_shift[i+1]-train_shift[i])
train_o_target.append(np.array(train_open)[i+1][29]+train_shift[i+1]-train_shift[i])
train_ohlc_target = np.column_stack((train_o_target, train_h_target, train_l_target, train_target[:-1]))
# obtendo test target para O,H,L
test_l_target = []
test_h_target = []
test_o_target = []
for i in range(test_shift.shape[0]-1):
test_l_target.append(np.array(test_low)[i+1][29]+test_shift[i+1]-test_shift[i])
test_h_target.append(np.array(test_high)[i+1][29]+test_shift[i+1]-test_shift[i])
test_o_target.append(np.array(test_open)[i+1][29]+test_shift[i+1]-test_shift[i])
test_ohlc_target = np.column_stack((test_o_target, test_h_target, test_l_target, test_target[:-1]))
#removendo ultimo elemento pq é perdido ao calcular os targets O,H e L
train_shift = train_shift[:-1]
train_target = train_target[:-1]
train_open = train_open[:-1]
train_high = train_high[:-1]
train_low = train_low[:-1]
train_close = train_close[:-1]
test_shift = test_shift[:-1]
test_target = test_target[:-1]
test_open = test_open[:-1]
test_high = test_high[:-1]
test_low = test_low[:-1]
test_close = test_close[:-1]
X_train, X_test, Y_train, Y_test = np.column_stack((train_open.values,train_high.values,train_low.values,train_close.values)), np.column_stack((test_open.values,test_high.values,test_low.values,test_close.values)), train_ohlc_target, test_ohlc_target #target OHLC
#X_train, X_test, Y_train, Y_test = np.column_stack((train_open.values,train_high.values,train_low.values,train_close.values)), np.column_stack((test_open.values,test_high.values,test_low.values,test_close.values)), np.array(train_target.values.reshape(train_target.size,1)), np.array(test_target.values.reshape(test_target.size,1))
X_trainp, X_testp, Y_trainp, Y_testp = X_train+train_shift.values.reshape(train_shift.size,1), X_test+test_shift.values.reshape(test_shift.size,1), Y_train+train_shift.values.reshape(train_shift.size,1), Y_test + test_shift.values.reshape(test_shift.size,1)
X_train = np.column_stack((X_train,Y_train)) # fica 120 + colunas target(1 ou 4) "features"
#apenas close
# X_train, X_test, Y_train, Y_test = np.array(train_close), np.array(test_close), np.array(train_target.values.reshape(train_target.size,1)), np.array(test_target.values.reshape(test_target.size,1))
# X_trainp, X_testp, Y_trainp, Y_testp = X_train+train_shift.values.reshape(train_shift.size,1), X_test+test_shift.values.reshape(test_shift.size,1), Y_train+train_shift.values.reshape(train_shift.size,1), Y_test + test_shift.values.reshape(test_shift.size,1)
# X_train = np.column_stack((X_train,Y_train))
def exec_time(start, msg):
end = time.time()
delta = end - start
if(delta > 60): print("Tempo: " + str(delta/60.0) + " min [" + msg + "]")
else: print("Tempo: " + str(int(delta)) + " s [" + msg + "]")
def generator_model(opt):
model = Sequential()
model.add(Dense(256, input_dim=120, kernel_initializer=RandomNormal(stddev=0.1)))
model.add(LeakyReLU(0.4))
model.add(Dense(512))
model.add(LeakyReLU(0.4))
model.add(Dense(1024))
model.add(LeakyReLU(0.4))
#model.add(Dense(121)) #output 120+previsao
model.add(Dense(4)) #output so previsao
model.add(Activation('tanh'))
model.compile(loss='binary_crossentropy', optimizer=opt)
return model
def discriminator_model(opt):
model = Sequential()
model.add(Dense(1024, input_dim=124, kernel_initializer=RandomNormal(stddev=0.1)))
model.add(LeakyReLU(0.4))
model.add(Dropout(0.5))
model.add(Dense(512))
model.add(LeakyReLU(0.4))
model.add(Dropout(0.5))
model.add(Dense(256))
model.add(LeakyReLU(0.4))
model.add(Dropout(0.5))
model.add(Dense(4, activation='sigmoid'))
model.compile(loss='binary_crossentropy', optimizer=opt)
return model
def gan_model(D, G, opt):
D.trainable = False
gan_input = Input(shape=(120,))
x = concatenate([gan_input, G(gan_input)])
gan_output = D(x)
gan = Model(gan_input, gan_output)
gan.compile(loss='binary_crossentropy', optimizer=opt)
return gan
def train(X_train, generator, discriminator, GAN, epochs=6000, verbose_step=250, batch_size=128, output_dir='output'):
print("*** Training", epochs, "epochs with batch size =", batch_size, "***")
times = []
d_lossses = []
g_losses = []
start_train = time.time()
for e in range(epochs+1):
start = time.time()
noise = np.random.normal(0, 1, size=[batch_size, 120])
input_data = X_train[np.random.randint(0, X_train.shape[0], size=2*batch_size)]
imageBatch = input_data[:int(input_data.shape[0]/2)]
#generator_data = np.delete(input_data[int(input_data.shape[0]/2):], -1, axis=1) # remove ultima coluna
generator_data = np.delete(input_data[int(input_data.shape[0]/2):], [120,121,122,123], axis=1) #remove ultimas 4 colunas
G_images = generator.predict(generator_data)
G_images = np.column_stack((generator_data,G_images))
X = np.concatenate([imageBatch, G_images])
#y = np.zeros(2*batch_size)
y = np.zeros((2*batch_size,4)) #OHLC target
y[:batch_size] = 0.9
discriminator.trainable = True
d_loss = discriminator.train_on_batch(X, y)
noise = np.random.normal(0, 1, size=[batch_size, 120])
#y = np.ones(batch_size)
y = np.ones((batch_size,4)) # OHLC target
discriminator.trainable = False
g_loss = GAN.train_on_batch(generator_data, y)
d_lossses.append(d_loss)
g_losses.append(g_loss)
times.append(time.time() - start)
plotGeneratedCandles("original_test", Y_testp[0:50,0], Y_testp[0:50,1], Y_testp[0:50,2], Y_testp[0:50,3], output_dir)
if(e % verbose_step == 0):
print(str(e) + ": d_loss =", d_loss, "| g_loss =", g_loss)
#trainPredict = generator.predict( np.delete(X_train, -1, axis=1))
trainPredict = generator.predict( np.delete(X_train, [120,121,122,123], axis=1))
testPredict = generator.predict(X_test)
new_predicted = testPredict+test_shift.values.reshape(test_shift.size,1)
new_train_predicted= trainPredict+train_shift.values.reshape(train_shift.size,1)
testScore = math.sqrt(mean_squared_error(new_predicted, Y_testp))
trainScore = math.sqrt(mean_squared_error(new_train_predicted, Y_trainp))
print("RMSE treino: %f"% trainScore)
print("RMSE test: %f" % testScore)
plotGeneratedCandles(e, new_predicted[0:50,0], new_predicted[0:50,1], new_predicted[0:50,2], new_predicted[0:50,3], output_dir)
#plotGeneratedImages(e, generator, output_dir)
exec_time(start_train, "Training")
generate_graphics(times, d_lossses, g_losses, output_dir)
def generate_graphics(times, d_lossses, g_losses, output_dir):
plt.close('all')
x = np.linspace(0, len(times), len(times))
plt.clf()
plt.title("GAN MNIST - Exec time per epoch")
plt.ylabel('seconds')
plt.xlabel('epoch')
plt.plot(x[1:], times[1:])
plt.savefig(os.path.join(output_dir, 'times.png'))
# plt.show()
plt.clf()
plt.title("GAN MNIST - D and G losses per epoch")
plt.ylabel('loss(binary crossentropy)')
plt.xlabel('epoch')
plt.plot(x, d_lossses, 'b-', label="D loss")
plt.plot(x, g_losses, 'g-', label="G loss")
plt.savefig(os.path.join(output_dir, 'losses.png'))
# plt.show()
def plotGeneratedImages(e, generator, output_dir, examples=100, dim=(10, 10), figsize=(10, 10)):
noise = np.random.normal(0, 1, size=[examples, 100])
generatedImages = generator.predict(noise)
generatedImages = generatedImages.reshape(examples, 28, 28)
plt.close('all')
plt.figure(figsize=figsize)
for i in range(generatedImages.shape[0]):
plt.subplot(dim[0], dim[1], i+1)
plt.imshow(generatedImages[i], interpolation='nearest', cmap='gray_r')
plt.axis('off')
plt.tight_layout()
plt.savefig(os.path.join(output_dir, str(e) + '.png'))
def plotGeneratedCandles(e, opens, highs, lows, closes, output_dir):
plt.close('all')
fig, ax = plt.subplots()
matplotlib.finance.candlestick2_ohlc(ax, opens, highs, lows, closes)
#plt.show()
plt.savefig(os.path.join(output_dir, str(e) + '.png'))
def main():
if(len(sys.argv) > 1):
folder = 'output_'+sys.argv[1]
if os.path.exists(folder):
shutil.rmtree(folder)
os.makedirs(folder)
opt = Adam(lr=0.0002, beta_1=0.5)
generator = generator_model(opt)
discriminator = discriminator_model(opt)
GAN = gan_model(discriminator, generator, opt)
train(X_train, generator, discriminator, GAN, output_dir=folder)
if __name__ == '__main__':
start = time.time()
main()
exec_time(start, "All")