cnn 一维时序数据_深度学习笔记31_创建一个CNN与RNN结合的预测模型

结合 CNN 和 RNN 来处理长序列

一维CNN在处理顺序敏感问题的弊端

一维卷积再处理序列文本的时候，它对时间序列的敏感度不是很强。因为这里通过上面的温度预测的模型来测试。 数据的准备参考前面的文章:

深度学习笔记26_RNN网络预测温度模型_数据准备 深度学习笔记27_RNN网络预测温度模型_搭建基准模型

# 构建模型
from keras.models import Sequential
from keras import layers
from keras.optimizers import RMSprop

model = Sequential()
model.add(layers.Conv1D(32,5,activation='relu',input_shape = (None,float_data.shape[-1])))
model.add(layers.MaxPool1D(3))
model.add(layers.Conv1D(32,5,activation = 'relu'))
model.add(layers.MaxPool1D(3))
model.add(layers.Conv1D(32,5,activation = 'relu'))

# 对于时序数据的全局最大池化
model.add(layers.GlobalMaxPooling1D())
model.add(layers.Dense(1)) 

model.summary()
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv1d_4 (Conv1D)            (None, None, 32)          2272      
_________________________________________________________________
max_pooling1d_3 (MaxPooling1 (None, None, 32)          0         
_________________________________________________________________
conv1d_5 (Conv1D)            (None, None, 32)          5152      
_________________________________________________________________
max_pooling1d_4 (MaxPooling1 (None, None, 32)          0         
_________________________________________________________________
conv1d_6 (Conv1D)            (None, None, 32)          5152      
_________________________________________________________________
global_max_pooling1d_2 (Glob (None, 32)                0         
_________________________________________________________________
dense_2 (Dense)              (None, 1)                 33        
=================================================================
Total params: 12,609
Trainable params: 12,609
Non-trainable params: 0
_________________________________________________________________
model.compile(optimizer=RMSprop(), loss='mae') 
history = model.fit_generator(train_gen, 
                              steps_per_epoch=500, 
                              epochs=20, 
                              validation_data=val_gen, 
                              validation_steps=1000)
Epoch 1/20
500/500 [==============================] - 14s 27ms/step - loss: 0.3628 - val_loss: 0.4323
Epoch 2/20
500/500 [==============================] - 13s 27ms/step - loss: 0.3398 - val_loss: 0.4596
Epoch 3/20
500/500 [