Reservoir Inflow Prediction by Employing Response Surface-Based Models Conjunction with Wavelet and Bootstrap Techniques

Table 4

Performances of numerous prediction models for 1 to 3 days lead time for testing data set 2010 (1 July to 31 September) of Chenab river basin.

Days	Model type	Model	NSE	RMSE (m^3/s)	MAE (m^3/s)	CP
1 d	Traditional	MLR	0.8426	0.4010	0.2082	0.7398
		FORS	0.8621	0.3780	0.1941	0.7719
		QRS	0.8913	0.3390	0.1952	0.8201
	Wavelet	WFORS	0.9761	0.2250	0.1470	0.9521
		WQRS	0.9796	0.2170	0.1354	0.9563
	Bootstrap	BFORS	0.9884	0.0214	0.0071	0.9981
		BQRS	0.9929	0.0079	0.0021	0.9961
2 d	Traditional	MLR	0.8429	0.4010	0.2061	0.7399
		FORS	0.8622	0.3770	0.1923	0.7719
		QRS	0.8913	0.3390	0.1942	0.8202
	Wavelet	WFORS	0.9779	0.2230	0.1437	0.9526
		WQRS	0.9799	0.2160	0.1328	0.9565
	Bootstrap	BFORS	0.9965	0.0128	0.0044	0.9977
		BQRS	0.9971	0.0082	0.0021	0.9988
3 d	Traditional	MLR	0.8430	0.4010	0.2059	0.7402
		FORS	0.8622	0.3770	0.1922	0.7719
		QRS	0.8914	0.3390	0.1936	0.8202
	Wavelet	WFORS	0.9782	0.2220	0.1415	0.9529
		WQRS	0.9800	0.2150	0.1316	0.9567
	Bootstrap	BFORS	0.9984	0.0122	0.0038	0.9984
		BQRS	0.9987	0.0049	0.0011	0.9994