请　[注册]　或　[登录]　 | 返回主站

量化交易吧 /  量化平台 帖子：3365817 新帖：24

深度学习模型 CNN+LSTM 预测收盘价

专门亏损发表于：5 月 10 日 02：51回复(1)

上一篇浏览量很大，感谢各位的关注！
能够在这里分享一些实验，一起领略 数据科学之美，也很开心。
以后，这个实验的模型会不断深化。
之后，也会分享一些 论文里 基于深度学习的时间序列预测模型。
数据由JQData本地量化金融数据支持
上一篇做了2个实验，预测黄金期货主力合约的收盘价。
实验2：
使?历史前5个时刻的 open close high low volume money
预测当前时刻的收盘价，
即 [None, 5, 6] => [None, 1] # None是 batch_size

这一篇对 第2个实验的模型 进行拓展，增加CNN层
因为 对每个样本是5行，6列的数据，二维数据，能够使用CNN进行特征提取

模型架构
输入层
CNN进行特征提取， 池化层 dropout
双向LSTM层
输出层
实验3.1.jpg
实验结果：是测试集的结果。test为测试集的真实收盘价，pred为模型预测的收盘价
实验3.png

import pandas as pd
import time, datetime
df_data_5minute=pd.read_csv('黄金主力5分钟数据.csv')
'''
或者使用JQdata
from jqdatasdk import *
#jqdata的账号密码
auth('邮箱：', 'jiaohiabin@ruc.edu.cn')
df_data_5minute= get_price('AU9999.XSGE',   start_date='2016-01-01', end_date='2018-01-01', frequency='5m')
'''

df_data_5minute.head()
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
Unnamed: 0 open close high low volume money
0 2016-01-04 09:05:00 226.70 226.65 226.85 226.45 5890.0 1.335146e+09
1 2016-01-04 09:10:00 226.75 226.50 226.75 226.40 2562.0 5.804133e+08
2 2016-01-04 09:15:00 226.45 226.45 226.60 226.40 1638.0 3.709666e+08
3 2016-01-04 09:20:00 226.45 226.25 226.50 226.20 3162.0 7.157891e+08
4 2016-01-04 09:25:00 226.25 226.25 226.30 226.20 1684.0 3.809907e+08 
df_data_5minute
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
open close high low volume money
0 226.70 226.65 226.85 226.45 5890.0 1.335146e+09
1 226.75 226.50 226.75 226.40 2562.0 5.804133e+08
2 226.45 226.45 226.60 226.40 1638.0 3.709666e+08
3 226.45 226.25 226.50 226.20 3162.0 7.157891e+08
4 226.25 226.25 226.30 226.20 1684.0 3.809907e+08
5 226.25 226.30 226.35 226.20 922.0 2.086313e+08
6 226.30 226.35 226.40 226.20 2476.0 5.603541e+08
7 226.30 226.45 226.45 226.25 2516.0 5.695246e+08
8 226.45 226.35 226.45 226.30 1344.0 3.042327e+08
9 226.30 226.30 226.35 226.20 1414.0 3.199363e+08
10 226.35 226.45 226.50 226.30 1610.0 3.645328e+08
11 226.45 226.40 226.50 226.40 972.0 2.200957e+08
12 226.40 226.50 226.55 226.35 2004.0 4.538166e+08
13 226.50 226.45 226.55 226.40 780.0 1.766423e+08
14 226.45 226.45 226.50 226.40 1530.0 3.464690e+08
15 226.55 226.45 226.65 226.45 2564.0 5.807784e+08
16 226.45 226.50 226.55 226.45 900.0 2.038475e+08
17 226.55 226.70 226.80 226.50 3008.0 6.817039e+08
18 226.70 226.65 226.85 226.60 2510.0 5.691306e+08
19 226.65 226.60 226.65 226.60 930.0 2.107595e+08
20 226.65 226.75 226.75 226.60 1184.0 2.683818e+08
21 226.75 226.65 226.75 226.60 1044.0 2.366603e+08
22 226.65 226.60 226.70 226.60 342.0 7.751130e+07
23 226.60 226.60 226.65 226.55 640.0 1.450196e+08
24 226.60 226.65 226.70 226.60 502.0 1.137778e+08
25 226.65 226.95 226.95 226.65 3222.0 7.308042e+08
26 226.90 226.90 226.95 226.80 1472.0 3.339398e+08
27 227.10 227.25 227.25 227.00 4894.0 1.111496e+09
28 227.25 227.55 227.60 227.20 5338.0 1.214103e+09
29 227.60 227.75 228.00 227.50 8612.0 1.961599e+09
... ... ... ... ... ... ...
53280 278.05 277.95 278.05 277.90 448.0 1.245318e+08
53281 277.90 277.95 278.00 277.90 506.0 1.406423e+08
53282 277.95 277.95 278.00 277.95 180.0 5.003790e+07
53283 277.95 278.00 278.05 277.95 936.0 2.602273e+08
53284 278.05 277.90 278.05 277.90 942.0 2.618281e+08
53285 277.85 277.90 277.95 277.85 518.0 1.439454e+08
53286 277.95 277.95 277.95 277.90 614.0 1.706443e+08
53287 277.90 277.90 277.95 277.85 1046.0 2.906776e+08
53288 277.95 277.90 277.95 277.90 206.0 5.725350e+07
53289 277.90 277.90 277.95 277.85 740.0 2.056435e+08
53290 277.90 277.85 277.90 277.85 200.0 5.557570e+07
53291 277.90 277.90 277.95 277.85 756.0 2.100840e+08
53292 277.90 278.00 278.00 277.90 490.0 1.362097e+08
53293 278.00 278.05 278.15 278.00 768.0 2.135675e+08
53294 278.10 278.15 278.15 278.05 252.0 7.008070e+07
53295 278.10 278.05 278.10 278.00 800.0 2.224430e+08
53296 278.00 278.00 278.05 277.95 184.0 5.115390e+07
53297 278.00 277.95 278.00 277.90 474.0 1.317464e+08
53298 277.95 277.95 277.95 277.90 334.0 9.282880e+07
53299 277.95 277.90 277.95 277.90 332.0 9.226560e+07
53300 277.90 277.95 277.95 277.90 672.0 1.867720e+08
53301 277.90 277.85 277.95 277.85 994.0 2.762458e+08
53302 277.90 277.90 277.95 277.85 352.0 9.781830e+07
53303 277.90 277.80 277.95 277.80 784.0 2.178426e+08
53304 277.85 277.80 277.85 277.75 920.0 2.555711e+08
53305 277.80 277.80 277.85 277.75 606.0 1.683349e+08
53306 277.80 277.85 277.85 277.80 560.0 1.555840e+08
53307 277.85 277.85 277.90 277.80 802.0 2.228271e+08
53308 277.85 277.75 277.90 277.75 1236.0 3.433855e+08
53309 277.80 277.80 277.90 277.70 1790.0 4.972797e+08

53310 rows × 6 columns

df_data_5minute.drop('Unnamed: 0', axis=1, inplace=True)
df_data_5minute
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
open close high low volume money
0 226.70 226.65 226.85 226.45 5890.0 1.335146e+09
1 226.75 226.50 226.75 226.40 2562.0 5.804133e+08
2 226.45 226.45 226.60 226.40 1638.0 3.709666e+08
3 226.45 226.25 226.50 226.20 3162.0 7.157891e+08
4 226.25 226.25 226.30 226.20 1684.0 3.809907e+08
5 226.25 226.30 226.35 226.20 922.0 2.086313e+08
6 226.30 226.35 226.40 226.20 2476.0 5.603541e+08
7 226.30 226.45 226.45 226.25 2516.0 5.695246e+08
8 226.45 226.35 226.45 226.30 1344.0 3.042327e+08
9 226.30 226.30 226.35 226.20 1414.0 3.199363e+08
10 226.35 226.45 226.50 226.30 1610.0 3.645328e+08
11 226.45 226.40 226.50 226.40 972.0 2.200957e+08
12 226.40 226.50 226.55 226.35 2004.0 4.538166e+08
13 226.50 226.45 226.55 226.40 780.0 1.766423e+08
14 226.45 226.45 226.50 226.40 1530.0 3.464690e+08
15 226.55 226.45 226.65 226.45 2564.0 5.807784e+08
16 226.45 226.50 226.55 226.45 900.0 2.038475e+08
17 226.55 226.70 226.80 226.50 3008.0 6.817039e+08
18 226.70 226.65 226.85 226.60 2510.0 5.691306e+08
19 226.65 226.60 226.65 226.60 930.0 2.107595e+08
20 226.65 226.75 226.75 226.60 1184.0 2.683818e+08
21 226.75 226.65 226.75 226.60 1044.0 2.366603e+08
22 226.65 226.60 226.70 226.60 342.0 7.751130e+07
23 226.60 226.60 226.65 226.55 640.0 1.450196e+08
24 226.60 226.65 226.70 226.60 502.0 1.137778e+08
25 226.65 226.95 226.95 226.65 3222.0 7.308042e+08
26 226.90 226.90 226.95 226.80 1472.0 3.339398e+08
27 227.10 227.25 227.25 227.00 4894.0 1.111496e+09
28 227.25 227.55 227.60 227.20 5338.0 1.214103e+09
29 227.60 227.75 228.00 227.50 8612.0 1.961599e+09
... ... ... ... ... ... ...
53280 278.05 277.95 278.05 277.90 448.0 1.245318e+08
53281 277.90 277.95 278.00 277.90 506.0 1.406423e+08
53282 277.95 277.95 278.00 277.95 180.0 5.003790e+07
53283 277.95 278.00 278.05 277.95 936.0 2.602273e+08
53284 278.05 277.90 278.05 277.90 942.0 2.618281e+08
53285 277.85 277.90 277.95 277.85 518.0 1.439454e+08
53286 277.95 277.95 277.95 277.90 614.0 1.706443e+08
53287 277.90 277.90 277.95 277.85 1046.0 2.906776e+08
53288 277.95 277.90 277.95 277.90 206.0 5.725350e+07
53289 277.90 277.90 277.95 277.85 740.0 2.056435e+08
53290 277.90 277.85 277.90 277.85 200.0 5.557570e+07
53291 277.90 277.90 277.95 277.85 756.0 2.100840e+08
53292 277.90 278.00 278.00 277.90 490.0 1.362097e+08
53293 278.00 278.05 278.15 278.00 768.0 2.135675e+08
53294 278.10 278.15 278.15 278.05 252.0 7.008070e+07
53295 278.10 278.05 278.10 278.00 800.0 2.224430e+08
53296 278.00 278.00 278.05 277.95 184.0 5.115390e+07
53297 278.00 277.95 278.00 277.90 474.0 1.317464e+08
53298 277.95 277.95 277.95 277.90 334.0 9.282880e+07
53299 277.95 277.90 277.95 277.90 332.0 9.226560e+07
53300 277.90 277.95 277.95 277.90 672.0 1.867720e+08
53301 277.90 277.85 277.95 277.85 994.0 2.762458e+08
53302 277.90 277.90 277.95 277.85 352.0 9.781830e+07
53303 277.90 277.80 277.95 277.80 784.0 2.178426e+08
53304 277.85 277.80 277.85 277.75 920.0 2.555711e+08
53305 277.80 277.80 277.85 277.75 606.0 1.683349e+08
53306 277.80 277.85 277.85 277.80 560.0 1.555840e+08
53307 277.85 277.85 277.90 277.80 802.0 2.228271e+08
53308 277.85 277.75 277.90 277.75 1236.0 3.433855e+08
53309 277.80 277.80 277.90 277.70 1790.0 4.972797e+08

53310 rows × 6 columns

df=df_data_5minute
close = df['close']
df.drop(labels=['close'], axis=1,inplace = True)
df.insert(0, 'close', close)
df
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
close open high low volume money
0 226.65 226.70 226.85 226.45 5890.0 1.335146e+09
1 226.50 226.75 226.75 226.40 2562.0 5.804133e+08
2 226.45 226.45 226.60 226.40 1638.0 3.709666e+08
3 226.25 226.45 226.50 226.20 3162.0 7.157891e+08
4 226.25 226.25 226.30 226.20 1684.0 3.809907e+08
5 226.30 226.25 226.35 226.20 922.0 2.086313e+08
6 226.35 226.30 226.40 226.20 2476.0 5.603541e+08
7 226.45 226.30 226.45 226.25 2516.0 5.695246e+08
8 226.35 226.45 226.45 226.30 1344.0 3.042327e+08
9 226.30 226.30 226.35 226.20 1414.0 3.199363e+08
10 226.45 226.35 226.50 226.30 1610.0 3.645328e+08
11 226.40 226.45 226.50 226.40 972.0 2.200957e+08
12 226.50 226.40 226.55 226.35 2004.0 4.538166e+08
13 226.45 226.50 226.55 226.40 780.0 1.766423e+08
14 226.45 226.45 226.50 226.40 1530.0 3.464690e+08
15 226.45 226.55 226.65 226.45 2564.0 5.807784e+08
16 226.50 226.45 226.55 226.45 900.0 2.038475e+08
17 226.70 226.55 226.80 226.50 3008.0 6.817039e+08
18 226.65 226.70 226.85 226.60 2510.0 5.691306e+08
19 226.60 226.65 226.65 226.60 930.0 2.107595e+08
20 226.75 226.65 226.75 226.60 1184.0 2.683818e+08
21 226.65 226.75 226.75 226.60 1044.0 2.366603e+08
22 226.60 226.65 226.70 226.60 342.0 7.751130e+07
23 226.60 226.60 226.65 226.55 640.0 1.450196e+08
24 226.65 226.60 226.70 226.60 502.0 1.137778e+08
25 226.95 226.65 226.95 226.65 3222.0 7.308042e+08
26 226.90 226.90 226.95 226.80 1472.0 3.339398e+08
27 227.25 227.10 227.25 227.00 4894.0 1.111496e+09
28 227.55 227.25 227.60 227.20 5338.0 1.214103e+09
29 227.75 227.60 228.00 227.50 8612.0 1.961599e+09
... ... ... ... ... ... ...
53280 277.95 278.05 278.05 277.90 448.0 1.245318e+08
53281 277.95 277.90 278.00 277.90 506.0 1.406423e+08
53282 277.95 277.95 278.00 277.95 180.0 5.003790e+07
53283 278.00 277.95 278.05 277.95 936.0 2.602273e+08
53284 277.90 278.05 278.05 277.90 942.0 2.618281e+08
53285 277.90 277.85 277.95 277.85 518.0 1.439454e+08
53286 277.95 277.95 277.95 277.90 614.0 1.706443e+08
53287 277.90 277.90 277.95 277.85 1046.0 2.906776e+08
53288 277.90 277.95 277.95 277.90 206.0 5.725350e+07
53289 277.90 277.90 277.95 277.85 740.0 2.056435e+08
53290 277.85 277.90 277.90 277.85 200.0 5.557570e+07
53291 277.90 277.90 277.95 277.85 756.0 2.100840e+08
53292 278.00 277.90 278.00 277.90 490.0 1.362097e+08
53293 278.05 278.00 278.15 278.00 768.0 2.135675e+08
53294 278.15 278.10 278.15 278.05 252.0 7.008070e+07
53295 278.05 278.10 278.10 278.00 800.0 2.224430e+08
53296 278.00 278.00 278.05 277.95 184.0 5.115390e+07
53297 277.95 278.00 278.00 277.90 474.0 1.317464e+08
53298 277.95 277.95 277.95 277.90 334.0 9.282880e+07
53299 277.90 277.95 277.95 277.90 332.0 9.226560e+07
53300 277.95 277.90 277.95 277.90 672.0 1.867720e+08
53301 277.85 277.90 277.95 277.85 994.0 2.762458e+08
53302 277.90 277.90 277.95 277.85 352.0 9.781830e+07
53303 277.80 277.90 277.95 277.80 784.0 2.178426e+08
53304 277.80 277.85 277.85 277.75 920.0 2.555711e+08
53305 277.80 277.80 277.85 277.75 606.0 1.683349e+08
53306 277.85 277.80 277.85 277.80 560.0 1.555840e+08
53307 277.85 277.85 277.90 277.80 802.0 2.228271e+08
53308 277.75 277.85 277.90 277.75 1236.0 3.433855e+08
53309 277.80 277.80 277.90 277.70 1790.0 4.972797e+08

53310 rows × 6 columns

data_train =df.iloc[:int(df.shape[0] * 0.7), :]
data_test = df.iloc[int(df.shape[0] * 0.7):, :]
print(data_train.shape, data_test.shape)

(37317, 6) (15993, 6)

# -*- coding: utf-8 -*-
import pandas as pd
import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
%matplotlib inline
from sklearn.preprocessing import MinMaxScaler
import time
scaler = MinMaxScaler(feature_range=(-1, 1))
scaler.fit(data_train)

MinMaxScaler(copy=True, feature_range=(-1, 1))
data_train = scaler.transform(data_train)
data_test = scaler.transform(data_test)

data_train

array([[-0.98877193, -0.98736842, -0.98459384, -0.99297259, -0.82504604,
        -0.85978547],
       [-0.99298246, -0.98596491, -0.98739496, -0.99437807, -0.92389948,
        -0.93904608],
       [-0.99438596, -0.99438596, -0.99159664, -0.99437807, -0.95134557,
        -0.96104178],
       ...,
       [ 0.61263158,  0.61824561,  0.61484594,  0.61349262, -0.90916652,
        -0.90885626],
       [ 0.61684211,  0.61403509,  0.61204482,  0.61630358, -0.94754352,
        -0.94737162],
       [ 0.6154386 ,  0.6154386 ,  0.61064426,  0.61349262, -0.94445435,
        -0.9442865 ]])
from keras.layers import Input, Dense, LSTM
from keras.models import Model
from keras.layers import *
from keras.models import *
from keras.optimizers import Adam

output_dim = 1
batch_size = 256
epochs = 60
seq_len = 5
hidden_size = 128


TIME_STEPS = 5
INPUT_DIM = 6

lstm_units = 64
X_train = np.array([data_train[i : i + seq_len, :] for i in range(data_train.shape[0] - seq_len)])
y_train = np.array([data_train[i + seq_len, 0] for i in range(data_train.shape[0]- seq_len)])
X_test = np.array([data_test[i : i + seq_len, :] for i in range(data_test.shape[0]- seq_len)])
y_test = np.array([data_test[i + seq_len, 0] for i in range(data_test.shape[0] - seq_len)])

print(X_train.shape, y_train.shape, X_test.shape, y_test.shape)

(37312, 5, 6) (37312,) (15988, 5, 6) (15988,)

inputs = Input(shape=(TIME_STEPS, INPUT_DIM))
#drop1 = Dropout(0.3)(inputs)

x = Conv1D(filters = 64, kernel_size = 1, activation = 'relu')(inputs)  #, padding = 'same'
#x = Conv1D(filters=128, kernel_size=5, activation='relu')(output1)#embedded_sequences
x = MaxPooling1D(pool_size = 5)(x)
x = Dropout(0.2)(x)

print(x.shape)

(?, 1, 64)

lstm_out = Bidirectional(LSTM(lstm_units, activation='relu'), name='bilstm')(x)
#lstm_out = LSTM(lstm_units,activation='relu')(x)
print(lstm_out.shape)

(?, 128)

output = Dense(1, activation='sigmoid')(lstm_out)
#output = Dense(10, activation='sigmoid')(drop2)

model = Model(inputs=inputs, outputs=output)
print(model.summary())

_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
input_7 (InputLayer)         (None, 5, 6)              0         
_________________________________________________________________
conv1d_7 (Conv1D)            (None, 5, 64)             448       
_________________________________________________________________
max_pooling1d_7 (MaxPooling1 (None, 1, 64)             0         
_________________________________________________________________
dropout_10 (Dropout)         (None, 1, 64)             0         
_________________________________________________________________
bilstm (Bidirectional)       (None, 128)               66048     
_________________________________________________________________
dense_6 (Dense)              (None, 1)                 129       
=================================================================
Total params: 66,625
Trainable params: 66,625
Non-trainable params: 0
_________________________________________________________________
None

model.compile(loss='mean_squared_error', optimizer='adam')
model.fit(X_train, y_train, epochs=epochs, batch_size=batch_size, shuffle=False)
y_pred = model.predict(X_test)
print('MSE Train loss:', model.evaluate(X_train, y_train, batch_size=batch_size))
print('MSE Test loss:', model.evaluate(X_test, y_test, batch_size=batch_size))
plt.plot(y_test, label='test')
plt.plot(y_pred, label='pred')
plt.legend()
plt.show()

Epoch 1/60
37312/37312 [==============================] - 4s 106us/step - loss: 0.1970
Epoch 2/60
37312/37312 [==============================] - 2s 43us/step - loss: 0.0618
Epoch 3/60
37312/37312 [==============================] - 2s 45us/step - loss: 0.0438
Epoch 4/60
37312/37312 [==============================] - 2s 48us/step - loss: 0.0434
Epoch 5/60
37312/37312 [==============================] - 1s 40us/step - loss: 0.0432
Epoch 6/60
37312/37312 [==============================] - 2s 43us/step - loss: 0.0429
Epoch 7/60
37312/37312 [==============================] - 1s 35us/step - loss: 0.0427
Epoch 8/60
37312/37312 [==============================] - 1s 33us/step - loss: 0.0425
Epoch 9/60
37312/37312 [==============================] - 1s 33us/step - loss: 0.0424
Epoch 10/60
37312/37312 [==============================] - 2s 49us/step - loss: 0.0422
Epoch 11/60
37312/37312 [==============================] - 1s 34us/step - loss: 0.0421
Epoch 12/60
37312/37312 [==============================] - 1s 32us/step - loss: 0.0419
Epoch 13/60
37312/37312 [==============================] - 1s 36us/step - loss: 0.0418
Epoch 14/60
37312/37312 [==============================] - 1s 37us/step - loss: 0.0417
Epoch 15/60
37312/37312 [==============================] - 1s 32us/step - loss: 0.0417
Epoch 16/60
37312/37312 [==============================] - 1s 32us/step - loss: 0.0416
Epoch 17/60
37312/37312 [==============================] - 1s 37us/step - loss: 0.0416
Epoch 18/60
37312/37312 [==============================] - 1s 36us/step - loss: 0.0415
Epoch 19/60
37312/37312 [==============================] - 1s 32us/step - loss: 0.0414
Epoch 20/60
37312/37312 [==============================] - 1s 32us/step - loss: 0.0414
Epoch 21/60
37312/37312 [==============================] - 1s 38us/step - loss: 0.0414
Epoch 22/60
37312/37312 [==============================] - 1s 36us/step - loss: 0.0414
Epoch 23/60
37312/37312 [==============================] - 1s 32us/step - loss: 0.0414
Epoch 24/60
37312/37312 [==============================] - 1s 32us/step - loss: 0.0414
Epoch 25/60
37312/37312 [==============================] - 1s 38us/step - loss: 0.0414
Epoch 26/60
37312/37312 [==============================] - 1s 37us/step - loss: 0.0414
Epoch 27/60
37312/37312 [==============================] - 1s 33us/step - loss: 0.0414
Epoch 28/60
37312/37312 [==============================] - 1s 36us/step - loss: 0.0414
Epoch 29/60
37312/37312 [==============================] - 1s 38us/step - loss: 0.0414
Epoch 30/60
37312/37312 [==============================] - 1s 35us/step - loss: 0.0413
Epoch 31/60
37312/37312 [==============================] - 1s 33us/step - loss: 0.0413
Epoch 32/60
37312/37312 [==============================] - 1s 36us/step - loss: 0.0413
Epoch 33/60
37312/37312 [==============================] - 2s 44us/step - loss: 0.0413
Epoch 34/60
37312/37312 [==============================] - 2s 41us/step - loss: 0.0413
Epoch 35/60
37312/37312 [==============================] - 1s 34us/step - loss: 0.0414
Epoch 36/60
37312/37312 [==============================] - 1s 39us/step - loss: 0.0413
Epoch 37/60
37312/37312 [==============================] - 1s 36us/step - loss: 0.0412
Epoch 38/60
37312/37312 [==============================] - 1s 32us/step - loss: 0.0412
Epoch 39/60
37312/37312 [==============================] - 1s 35us/step - loss: 0.0412
Epoch 40/60
37312/37312 [==============================] - 1s 39us/step - loss: 0.0412
Epoch 41/60
37312/37312 [==============================] - 1s 34us/step - loss: 0.0413
Epoch 42/60
37312/37312 [==============================] - 1s 33us/step - loss: 0.0412
Epoch 43/60
37312/37312 [==============================] - 1s 37us/step - loss: 0.0412
Epoch 44/60
37312/37312 [==============================] - 1s 38us/step - loss: 0.0412
Epoch 45/60
37312/37312 [==============================] - 1s 33us/step - loss: 0.0412
Epoch 46/60
37312/37312 [==============================] - 1s 34us/step - loss: 0.0411
Epoch 47/60
37312/37312 [==============================] - 1s 39us/step - loss: 0.0412
Epoch 48/60
37312/37312 [==============================] - 1s 38us/step - loss: 0.0411
Epoch 49/60
37312/37312 [==============================] - 1s 33us/step - loss: 0.0412
Epoch 50/60
37312/37312 [==============================] - 1s 37us/step - loss: 0.0411
Epoch 51/60
37312/37312 [==============================] - 2s 43us/step - loss: 0.0411
Epoch 52/60
37312/37312 [==============================] - 1s 37us/step - loss: 0.0411
Epoch 53/60
37312/37312 [==============================] - 1s 33us/step - loss: 0.0412
Epoch 54/60
37312/37312 [==============================] - 1s 37us/step - loss: 0.0410
Epoch 55/60
37312/37312 [==============================] - 1s 39us/step - loss: 0.0411
Epoch 56/60
37312/37312 [==============================] - 1s 34us/step - loss: 0.0411
Epoch 57/60
37312/37312 [==============================] - 1s 33us/step - loss: 0.0411
Epoch 58/60
37312/37312 [==============================] - 2s 47us/step - loss: 0.0410
Epoch 59/60
37312/37312 [==============================] - 2s 41us/step - loss: 0.0411
Epoch 60/60
37312/37312 [==============================] - 1s 35us/step - loss: 0.0410
37312/37312 [==============================] - 1s 24us/step
MSE Train loss: 0.041352386607933875
15988/15988 [==============================] - 0s 15us/step
MSE Test loss: 0.0003156892797136216

随着训轮数（epoch）的增加，误差（loss）不断减小 loss: 0.0410左右

       

全部回复

0/140

达人推荐

量化课程

    移动端课程