A Hybrid Deep Learning Framework for Network Flow Forecasting of Power Grid Enterprise

Table 2

Single-step forecasting (one hour ahead) results and error comparison among different forecasting models.

Models	Inflow				Outflow
	MAE	MSE	R2	adjR2	MAE	MSE	R2	adjR2
GRU	32309057	3.45e15	0.7761	0.7729	38087998	4.4e15	0.7542	0.7507
VMD-GRU	30587150	1.19e15	0.9244	0.9233	41196151	2.26e15	0.8796	0.8779
Xgboost	49144727	5.52e15	0.6644	0.6597	97683111	2.91e16	0.4462	0.4384
VMD-xgboost	9026284	1.36e14	0.9915	0.9914	11215930	2.08e14	0.9886	0.9885
VMD-VMD-xgboost	9034632	1.36e14	0.9915	0.9914	11229317	2.09e14	0.9886	0.9884
VMD-xgboost-adjustment	8525637	1.19e14	0.9928	0.9927	10359168	1.77e14	0.9908	0.9906
EMD-xgboost	24017235	1.41e15	0.9195	0.9184	39651555	2.8e15	0.8796	0.8779
Proposed	6632489	7.55e13	0.9955	0.9954	8139553	1.13e14	0.9942	0.9941