A Comparative Study of VMD-Based Hybrid Forecasting Model for Nonstationary Daily Streamflow Time Series

Table 8

The performance of the different prediction models during the testing period for the Wushan and Weijiabao hydrological stations in the Wei River Basin.

Stations	Models	NSE	RMSE (m^3/s)	MAE (m^3/s)	Time cost (s)
Wushan	BPNN	0.8132	7.87	3.57	5.33
	EEMD-BPNN	0.9346	4.65	2.00	176.23
	VMD-ARMA	0.9547	3.88	1.86	1941.42
	VMD-GBRT	0.9399	4.46	2.09	158.23
	VMD-SVR	0.9395	4.47	2.00	4330.57
	VMD-BPNN	0.9566	3.79	1.84	54.45
Weijiabao	BPNN	0.7698	68.86	24.79	4.37
	EEMD-BPNN	0.9236	39.66	13.96	198.88
	VMD-ARMA	0.9862	16.84	7.99	3211.78
	VMD-GBRT	0.9750	22.70	10.21	281.34
	VMD-SVR	0.9857	17.18	8.13	4673.64
	VMD-BPNN	0.9865	16.80	7.95	123.13