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

Table 5

Performance of the SVR model during the training period at the Weijiaobao station.

Sequence	SVR parameters			Training
	C			RMSE (m^3/s)	NSE
IMF1	25	0.0000001	1	0.227	0.999992
IMF2	25	0.0000001	1	0.335	0.999954
IMF3	24.346	0.0000001	0.377	0.432	0.999867
IMF4	24.545	0.0000001	0.433	0.535	0.999728
IMF5	25	0.0000001	0.366	0.777	0.998884
IMF6	25	0.0000001	1	0.706	0.999041
IMF7	23.428	0.0002089	0.465	0.776	0.996972
IMF8	21.469	0.0005338	0.031	0.910	0.995910
IMF9	14.231	0.0000001	1	0.690	0.997577
IMF10	19.407	0.0000001	0.652	0.497	0.998274
IMF11	25	0.0000001	1	0.487	0.997650