Forecasting China’s per Capita Living Energy Consumption by Employing a Novel DGM (1, 1, t^α) Model with Fractional Order Accumulation

Table 4

Performance evaluation indicators of the nine prediction models.

Fitting	GM	DGM (1)	FDGM	NGM	NDGM	FNDGM	ARGM	DGM (2)	FDGM
MAPE	3.5060	3.5080	1.1360	1.2250	1.1970	1.2190	1.563	1.1310	1.0720
RMSPE	4.5960	4.6100	1.4330	1.4810	1.4670	1.4750	1.8540	1.3940	1.4010
MAE	764.1110	764.0000	296.2220	315.1110	311.2220	306.0000	378.7780	293.7780	257.6670
MSE	931.7600	931.8150	378.2150	392.9960	393.1800	378.6850	463.7160	396.0050	347.5530
IA	0.9857	0.9857	0.9978	0.9976	0.9976	0.9978	0.9967	0.9979	0.9981
R	0.9814	0.9814	0.9970	0.9967	0.9967	0.9969	0.9959	0.9971	0.9974
Prediction	GM	DGM (1)	FDGM	NGM	NDGM	FNDGM	ARGM	DGM (2)	FDGM (1,1, t^α)
MAPE	2.9880	2.9640	11.6790	14.0770	14.2990	10.7190	10.2540	12.1170	2.3850
RMSPE	3.2580	3.2310	12.7210	15.2380	15.4650	11.6640	11.4410	13.1280	2.7860
MAE	1108.170	1099.000	4395.330	5289.670	5372.000	4032.830	3881.500	4554.000	899.000
MSE	1224.470	1213.720	4948.640	5914.460	6000.425	4536.850	4481.360	5098.680	1094.950
IA	0.9728	0.9732	0.5855	0.5177	0.5126	0.6206	0.6131	0.5767	0.9808
R	0.9961	0.9961	0.9869	0.9789	0.9783	0.9894	0.9845	0.9868	0.9968

Note. DGM (1) = DGM (1, 1); DGM (2) = DGM ((1, 1, t^α).