Table 2.

Detailed parameters of the baseline algorithms.

Name	Architecture
DT	criterion = gini
RF	n_estimators = 50
SVM	Kernel = rbf
CNN	C(64)- C(128)-GAP- FC(8)-Softmax
LSTM	LSTM(128)-FC(8)-Softmax
CNN + LSTM TCMH	C(64)-LSTM(128)-FC(8)-Softmax TCN(32)-MHA(5)-CNN(64)-FC(8)-Softmax

Note: FC denotes fully connected layer; C denotes convolutional neural network; GAP denotes global average pooling.