Developing Deep Survival Model for Remaining Useful Life Estimation Based on Convolutional and Long Short-Term Memory Neural Networks

Table 5

Results of FD003 with two analyses and different TW.

	Loss function (regression)				Loss function (Weibull)
TW	MAE	RMSE	^2	Score	MAE	RMSE	^2	Score
20	14.41	20.20	0.79	2,437	33.45	49.28	0.52	4,598,400
30	11.79	17.12	0.85	1,700	24.94	37.87	0.65	139,633
40	11.01	16.02	0.83	1091	15.86	24.33	0.88	9,330
50	11.24	15.41	—	679	14.45	21.82	0.90	8,597
60	10.80	14.62	0.86	325	14.69	23.57	0.87	32,361
70	12.08	15.94	0.80	401	14.11	22.73	0.88	22,581
80	11.87	16.54	0.82	379	13.69	23.01	0.87	30,087
90	12.25	17.25	0.79	353	14.00	23.42	0.86	39,985
100	11.88	17.53	0.78	366	14.60	25.52	0.83	30,360
110	12.58	19.41	0.76	463	12.21	20.98	0.89	11,536
120	13.45	21.20	—	625	9.81	15.55	0.93	672