Multivariate CNN-LSTM Model for Multiple Parallel Financial Time-Series Prediction

Table 2

Parameters configuration for the proposed multivariate CNN-LSTM model.

Parameters	Values	Layer
Convolutional layer filters	64	CNN
Convolutional layer kernel_size	2	CNN
Convolutional layer activation function	Relu	CNN
Convolutional layer padding	Valid	CNN
Pooling layer pool_size	2	CNN
Number of LSTM layer 1 hidden unit	100	LSTM
Number of LSTM layer 2 hidden unit	100	LSTM
LSTM layer 1 and 2 activation function	Relu	LSTM
Time step	4	LSTM
Optimizer	Adam	Model fitting
Loss function	MAE	Model fitting
Epochs	1000	Model fitting