A Hybrid Model Based on Ensemble Empirical Mode Decomposition and Fruit Fly Optimization Algorithm for Wind Speed Forecasting

Table 7

The forecasting results of model selection among the FOARBF, FOAGRNN, and FOASVR in winter.

Components	Error criteria	FOARBF	FOAGRNN	FOASVR
IMF2	MAE	0.1980	0.1564	0.0736
	RMSE	0.2516	0.1936	0.0954
	IA	0.8183	0.8868	0.9802
IMF3	MAE	0.1191	0.0475	0.0286
	RMSE	0.1494	0.0617	0.0351
	IA	0.9481	0.9907	0.9972
IMF4	MAE	0.1802	0.0631	0.0173
	RMSE	0.2120	0.0775	0.0212
	IA	0.9224	0.9921	0.9994
IMF5	MAE	0.0399	0.0661	0.0928
	RMSE	0.0491	0.0722	0.1013
	IA	0.9982	0.9958	0.9921
IMF6	MAE	0.1175	0.0144	0.1348
	RMSE	0.1207	0.0162	0.1424
	IA	0.9902	0.9998	0.9853
IMF7	MAE	0.3543	0.0066	0.0571
	RMSE	0.4067	0.0066	0.0889
	IA	0.4432	0.9998	0.9394
	MAE	0.0775	0.0024	0.0086
	RMSE	0.0810	0.0025	0.0101
	IA	0.3960	0.9982	0.9655