Improving Accuracy of River Flow Forecasting Using LSSVR with Gravitational Search Algorithm

Table 3

Comparison of the HLGSA, MARS, M5RT, and CMLR methods: Shyok catchment.

Indexes	Cross validation data sets	Input combinations	Methods
			HLGSA	MARS	M5RT	CMLR
	DS1	IC1	168.82	181.90	208.64	352.90
		IC2	145.96	156.39	171.05	209.44
		IC3	123.69	152.65	155.46	173.55
	DS2	IC1	176.70	214.12	251.98	328.46
		IC2	151.87	178.95	209.81	224.69
		IC3	135.67	176.82	177.60	182.70
	DS3	IC1	163.33	178.43	200.20	300.36
		IC2	138.85	149.03	149.96	201.21
		IC3	109.20	128.38	141.13	151.21
	DS4	IC1	209.95	232.83	264.04	371.26
		IC2	185.74	193.75	218.38	257.42
		IC3	170.68	181.39	189.40	195.10
	DS1	IC1	102.17	110.63	109.72	220.14
		IC2	78.63	92.31	107.78	135.56
		IC3	67.52	87.96	100.71	94.33
	DS2	IC1	106.96	128.49	113.52	187.27
		IC2	81.68	94.76	100.12	121.25
		IC3	87.98	93.26	93.68	96.48
	DS3	IC1	90.56	105.66	105.78	181.46
		IC2	66.59	74.29	97.34	116.65
		IC3	55.75	69.52	77.27	86.70
	DS4	IC1	120.19	132.56	137.95	224.07
		IC2	104.50	107.79	112.15	147.17
		IC3	88.99	102.18	103.53	105.04
	DS1	IC1	0.861	0.840	0.810	0.597
		IC2	0.904	0.878	0.841	0.786
		IC3	0.909	0.895	0.888	0.864
	DS2	IC1	0.872	0.816	0.757	0.616
		IC2	0.886	0.869	0.851	0.794
		IC3	0.905	0.879	0.865	0.846
	DS3	IC1	0.920	0.859	0.816	0.621
		IC2	0.929	0.890	0.876	0.841
		IC3	0.947	0.917	0.891	0.876
	DS4	IC1	0.811	0.781	0.719	0.524
		IC2	0.858	0.867	0.802	0.726
		IC3	0.920	0.879	0.861	0.820