Seismic Events Prediction Using Deep Temporal Convolution Networks

Table 3

Seven compared methods performance on AAIA′16 seismic dataset.

Category	Methods	RMSE	MAE	Running time (s)
Traditional time series prediction	ARIMA	2.038	1.602	800.3
Classic machine learning method	SVR	0.698	0.586	315.6
	RF	0.455	0.405	105.5
Normal recurrent network	LSTM	0.366	0.284	1496.2
Normal temporal CNN network	TCNN	0.198	0.159	1347.7
The proposed method	DCTCNN		0.070	4729.3
	CNN-LSTM	0.011	0.109	2302.4