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

Table 4

Performance of the ARMA model during the training period at the Weijiabao station.

Sequence	ARMA parameters		Training
	p	q	RMSE (m^3/s)	NSE
IMF1	19	18	0.145	0.999997
IMF2	13	19	0.196	0.999987
IMF3	20	18	0.251	0.999929
IMF4	6	20	0.334	0.999874
IMF5	4	19	0.442	0.999604
IMF6	20	19	0.508	0.999132
IMF7	7	20	0.563	0.999325
IMF8	18	17	0.546	0.998371
IMF9	19	19	0.496	0.996429
IMF10	20	20	0.415	0.995971
IMF11	20	20	0.314	0.997730