Seismic Events Prediction Using Deep Temporal Convolution Networks

Algorithm 2

Pseudo code of CNN-LSTM.

	Input: Multivariable time series (training set and testing set)
	Output: Predicted values , trained network weights:
	Parameters:
	Kernel size: k
	Training iterations: M
	Training batch size: B
	Learning rate:
	1: Load training set and testing set from
	2: Randomly initialize weight
	3: Begin Training
	4: for [1, M] do
	5: Forward passing as equations (12)–(17)
	6: Calculate loss as equation (18)
	7: Calculate gradients of weights
	8: Backpropagation and update
	9: End Training
	10: Begin Testing: calculate predicted values with testing set
	11: return predicted values ,