A Compound Structure for Wind Speed Forecasting Using MKLSSVM with Feature Selection and Parameter Optimization

Table 9

Forecasting results by EEMD-HGSA-LSSVM based on different decomposition method.

Data set	Models	RMSE (m/s)	MAPE ()	MAE (m/s)
A	Model 4	0.5119	9.9989	0.4785
	Model 5	0.3979	7.8679	0.3514
	Model 6	0.3843	7.883	0.3596
	Model 7	0.2787	5.457	0.2558
	Model 4	0.5406	13.3218	0.5138
	Model 5	0.4168	9.9403	0.3927
	Model 6	0.4029	9.658	0.3824
	Model 7	0.2856	6.4077	0.2538
	Model 4	0.5181	6.5663	0.4985
	Model 5	0.409	4.5731	0.3456
	Model 6	0.3882	4.8105	0.3642
	Model 7	0.2691	3.232	0.2392
	Model 4	0.5406	8.7821	0.5064
	Model 5	0.4025	6.3819	0.3701
	Model 6	0.3908	6.3498	0.3687
	Model 7	0.2903	4.5088	0.2607

Model 4 denotes HGSA-MKLSSVM.
Models 5, 6, and 7 denote HGSA-MKLSSVM with WT, EMD, and EEMD, respectively.