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

Table 5

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

Components	Error criteria	FOARBF	FOAGRNN	FOASVR
IMF2	MAE	0.0617	0.1521	0.0807
	RMSE	0.0756	0.1857	0.1161
	IA	0.9883	0.9206	0.9718
IMF3	MAE	0.1470	0.0874	0.0670
	RMSE	0.1919	0.1021	0.0772
	IA	0.9296	0.9825	0.9904
IMF4	MAE	0.2023	0.0419	0.0681
	RMSE	0.2355	0.0513	0.0759
	IA	0.9387	0.9978	0.9952
IMF5	MAE	0.0571	0.0397	0.0228
	RMSE	0.0656	0.0491	0.0256
	IA	0.9670	0.9824	0.9949
IMF6	MAE	0.0136	0.4352	0.0904
	RMSE	0.0148	0.4580	0.1027
	IA	0.9977	0.3439	0.8650
IMF7	MAE	0.0024	0.0022	0.0024
	RMSE	0.0025	0.0026	0.0027
	IA	0.9871	0.9864	0.9849
	MAE	0.0501	0.0366	0.0672
	RMSE	0.0595	0.0376	0.0701
	IA	0.9026	0.9682	0.8874