Integrating Feature Engineering with Deep Learning to Conduct Diagnostic and Predictive Analytics for Turbofan Engines

Table 10

Forecasting errors using PCA (principal component analysis).

Training	RF	XGB	SVM	DNN	RNN	LSTM	GRU	CNN
RMSE	18.68	18.47	20.13	20.44	20.25	20.74	19.9	21.47
MAE	13.66	14.06	14.24	16.16	15.93	16.16	15.51	17.42
MAAPE	20.9	20.69	22.76	22.81	22.96	23.31	23.2	23.07
Time (s)	59	6564	190	989	3461	9897	9344	2306
Testing	RF	XGB	SVM	DNN	RNN	LSTM	GRU	CNN
RMSE	17.58	17.14	20.2	17.8	17.72	18.71	17.76	17.6
MAE	12.36	12.05	14.09	12.46	12.3	13.09	12.25	12.45
MAAPE	20.64	19.93	23.54	19.59	19.32	20.92	19.68	19.92
Time (s)	0.47	3.14	2.22	0.29	0.33	0.58	0.34	0.31