Table 6.

The optimal hyper-parameter settings.

Module III and IV	Parameter	Value
K-means clustering algorithm	K	10
LSTM networks	Number of Hidden Layers	2
	Hidden layer Units	32/64
	Batch Size	10
	Epochs	100
	Learning Rate	0.001
	Optimizer	Adam
	Loss Function	Sparse_categorical_crossentropy
	Validation Frequency	1