-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathlstm_ae_snp500.py
More file actions
235 lines (200 loc) · 12.2 KB
/
lstm_ae_snp500.py
File metadata and controls
235 lines (200 loc) · 12.2 KB
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
import os
from functools import partial
from torch import nn
from Architectures.lstm_autoencoder import ToyAutoEncoder, SP500AutoEncoder, SP500AutoEncoder_prediction
from Utils.data_utils import DataUtils
from Utils.parameters_tune import ParameterTuning
from Utils.training_utils import TrainingUtils
from Utils.visualization_utils import VisualizationUtils
import torch
from torch.utils.tensorboard import SummaryWriter
import argparse
import numpy as np
writer = SummaryWriter()
parser = argparse.ArgumentParser(description='lstm_ae_snp500')
parser.add_argument('--batch-size', type=int, default=40, metavar='N',
help='input batch size for training (default: 128)')
parser.add_argument('--epochs', type=int, default=30, metavar='N', # 100,150
help='number of epochs to train (default: 10)')
parser.add_argument('--lstm-layers-size', type=int, default=3, metavar='N',
help='lstm layers number, default 3')
parser.add_argument('--lstm-dropout', type=int, default=0.2, metavar='N',
help='lstm layers number, default 0')
parser.add_argument('--optimizer', type=str, default="adam", metavar='N',
help='optimizer, default adam')
parser.add_argument('--load', type=bool, default=True, metavar='N',
help='To load or create new data, default True')
parser.add_argument('--decoder-output-size', type=int, default=256, metavar='N',
help='LSTM size at the end, default 256')
parser.add_argument('--folds', type=int, default=1, metavar='N',
help='k for k fold')
args = parser.parse_args()
print(torch.cuda.get_device_name(0))
def plot_stock_high_prices(path, plots_suffix):
amazon_daily_max, googl_daily_max = DataUtils.load_snp500_amzn_google_daily_max(path)
VisualizationUtils.plot_df_columns(amazon_daily_max,
"date",
"high",
"Amazon \nTime vs Daily Maximum",
"Time",
"Daily high",
os.path.join(plots_suffix, "amazon"))
VisualizationUtils.plot_df_columns(googl_daily_max,
"date",
"high",
"Google \nTime vs Daily Maximum",
"Time",
"Daily high",
os.path.join(plots_suffix, "google"))
def snp500_reconstruct(data_tensor, config, device, plots_suffix, n_part):
train_idxs, test_idxs = DataUtils.create_random_train_test_indices_split(data_tensor.shape[0], 0.85, 0.15)
data_gen = DataUtils.generate_random_split(data_tensor[train_idxs], args.folds, 0.9, 0.1)
test_loader = DataUtils.create_data_loader(data_tensor[test_idxs].unsqueeze(2), args.batch_size)
criterion = nn.MSELoss()
tune = ParameterTuning()
tune.kfold_run(train_func=partial(TrainingUtils.kfold_train,
# auto_encoder_init=ToyAutoEncoder,
auto_encoder_init=SP500AutoEncoder,
lr=config["lr"],
hidden_size=config["hidden_size"],
input_size=1,
input_seq_size=data_tensor.shape[1], # 1007
batch_size=args.batch_size,
criterion=criterion,
optimizer=args.optimizer,
lstm_layers_size=args.lstm_layers_size,
decoder_output_size=args.decoder_output_size,
epochs=args.epochs,
training_iteration=TrainingUtils.training_iteration,
validation=TrainingUtils.validation,
device=device),
test_func=partial(TrainingUtils.test_accuracy,
criterion=criterion,
test_loader=test_loader,
device=device),
data_tensor=data_tensor[train_idxs].unsqueeze(2),
data_generator=data_gen,
batch_size=args.batch_size,
config=config)
test_loader = DataUtils.create_data_loader(data_tensor[test_idxs].unsqueeze(2),
min(len(test_idxs), args.batch_size))
test_input = next(iter(test_loader))
test_input = test_input.to(device)
reconstructed = tune.best_model(test_input)
tune.plot_all_results(plots_suffix, is_accuracy=False, is_gridsearch=False, n_part=n_part)
test_input = np.squeeze(test_input.cpu().detach().numpy(), 2)
reconstructed = np.squeeze(reconstructed.cpu().detach().numpy(), 2)
return test_input, reconstructed
def reconstruct():
data_dir = os.path.join("data", "SP 500 Stock Prices 2014-2017.csv")
plots_suffix = os.path.join("plots", "snp500", "part_II")
config = {"hidden_size": 256,
"lr": 0.001,
"grad_clip": None}
device = "cuda:0" if torch.cuda.is_available() else "cpu"
data_tensor, train_stocks_names = DataUtils.load_snp500(data_dir, args.batch_size, 10)
test_input, reconstructed = snp500_reconstruct(data_tensor[0],
config,
device,
plots_suffix,
"_part"+str(0)+"_")
for i in range(len(data_tensor) - 1):
sub_test_input, sub_reconstructed = snp500_reconstruct(data_tensor[i + 1],
config,
device,
plots_suffix,
"_part"+str(i + 1)+"_")
test_input = np.concatenate((test_input, sub_test_input), axis=1)
reconstructed = np.concatenate((reconstructed, sub_reconstructed), axis=1)
VisualizationUtils.plot_reconstruct(reconstructed,
test_input,
3,
os.path.join(plots_suffix, "reconstruct"),
"Reconstructed vs Original",
["Origin", "Reconstructed"])
def snp500_prediction(data_tensor, config, device, plots_suffix, n_part):
train_idxs, test_idxs = DataUtils.create_random_train_test_indices_split(data_tensor.shape[0], 0.85, 0.15)
data_gen = DataUtils.generate_random_split(data_tensor[train_idxs], args.folds, 0.8, 0.2)
test_loader = DataUtils.create_data_loader(data_tensor[test_idxs].unsqueeze(2), args.batch_size)
criterion = nn.MSELoss()
tune = ParameterTuning()
tune.kfold_run(train_func=partial(TrainingUtils.kfold_train,
# auto_encoder_init=ToyAutoEncoder,
auto_encoder_init=SP500AutoEncoder_prediction,
lr=config["lr"],
hidden_size=config["hidden_size"],
input_size=2,
input_seq_size=int(data_tensor.shape[1] / 2), # 1007
batch_size=args.batch_size,
criterion=criterion,
optimizer=args.optimizer,
lstm_layers_size=args.lstm_layers_size,
decoder_output_size=args.decoder_output_size,
epochs=args.epochs,
training_iteration=TrainingUtils.prediction_training_iteration,
validation=TrainingUtils.prediction_validation,
device=device),
test_func=partial(TrainingUtils.prediction_test,
criterion=criterion,
test_loader=test_loader,
device=device),
data_tensor=data_tensor[train_idxs].unsqueeze(2),
data_generator=data_gen,
batch_size=args.batch_size,
config=config)
test_loader = DataUtils.create_data_loader(data_tensor[test_idxs].unsqueeze(2),
min(len(test_idxs), args.batch_size))
test_input = next(iter(test_loader))
test_input = test_input.to(device)
b, r, c = test_input.shape
test_input = test_input.view(b, int(r / 2), 2)
reconstruct, predict = tune.best_model(test_input)
tune.plot_all_results(plots_suffix, is_accuracy=False, is_gridsearch=False, n_part=n_part)
original_seq_first_day = test_input[:, :, 0].cpu().detach().numpy()
original_seq_second_day = test_input[:, :, 1].cpu().detach().numpy()
reconstruct = reconstruct.detach().cpu().numpy()
predict = predict.detach().cpu().numpy()
return original_seq_first_day, reconstruct, original_seq_second_day, predict
def prediction():
data_dir = os.path.join("data", "SP 500 Stock Prices 2014-2017.csv")
plots_suffix = os.path.join("plots", "snp500", "part_III")
config = {"hidden_size": 256,
"lr": 0.001,
"grad_clip": None}
device = "cuda:0" if torch.cuda.is_available() else "cpu"
data_tensor, train_stocks_names = DataUtils.load_snp500_double_input(data_dir, args.batch_size, 10)
seq_first_day, reconstruct, seq_second_day, predict = snp500_prediction(data_tensor[0],
config,
device,
plots_suffix,
"_part"+str(0)+"_")
for i in range(len(data_tensor) - 1):
sub_seq_first_day, sub_reconstruct, sub_seq_second_day, sub_predict = snp500_prediction(data_tensor[i + 1],
config,
device,
plots_suffix,
"_part"+str(i + 1)+"_")
seq_first_day = np.concatenate((seq_first_day, sub_seq_first_day), axis=1)
reconstruct = np.concatenate((reconstruct, sub_reconstruct), axis=1)
seq_second_day = np.concatenate((seq_second_day, sub_seq_second_day), axis=1)
predict = np.concatenate((predict, sub_predict), axis=1)
VisualizationUtils.plot_reconstruct(reconstruct,
seq_first_day,
3,
os.path.join(plots_suffix, "Reconstructed"),
"Reconstructed vs Original",
["Origin", "Reconstructed"])
VisualizationUtils.plot_reconstruct(predict,
seq_second_day,
3,
os.path.join(plots_suffix, "predict"),
"Prediction vs Original",
["Origin", "Prediction"])
def plot_stocks():
path = os.path.join("data", "SP 500 Stock Prices 2014-2017.csv")
plots_suffix = os.path.join("plots", "snp500", "part_I")
plot_stock_high_prices(path, plots_suffix)
if __name__ == "__main__":
plot_stocks() #3.1
reconstruct() #3.2
prediction() #3.3