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

Table 6

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

Components	Error criteria	FOARBF	FOAGRNN	FOASVR
IMF2	MAE	0.1206	0.2141	0.0884
	RMSE	0.1647	0.2888	0.1049
	IA	0.9640	0.8839	0.9874
IMF3	MAE	0.0755	0.0662	0.0435
	RMSE	0.0984	0.0838	0.0535
	IA	0.9798	0.9849	0.9940
IMF4	MAE	0.2501	0.0549	0.0247
	RMSE	0.2873	0.0639	0.0305
	IA	0.9396	0.9974	0.9994
IMF5	MAE	0.0488	0.1090	0.0722
	RMSE	0.0553	0.1252	0.0777
	IA	0.9996	0.9977	0.9991
IMF6	MAE	0.0745	0.0677	0.0275
	RMSE	0.0999	0.0685	0.0279
	IA	0.9761	0.9909	0.9985
IMF7	MAE	0.0217	0.0194	0.0273
	RMSE	0.0244	0.0196	0.0273
	IA	0.9852	0.9889	0.9773
	MAE	0.1185	0.0756	0.0055
	RMSE	0.1281	0.0803	0.0068
	IA	0.2589	0.4183	0.9875