A Novel Hybrid Model for Short-Term Wind Speed Forecasting Based on Twice Decomposition, PSR, and IMVO-ELM

Table 2

The average evaluation metrics of training by different methods with 20 times’ independent calculations.

Datasets	Methods	Evaluation metrics of training
		MAPE (%)	MAE (m/s)	RMSE (m/s)
A	ELM	15.278	0.865	1.143
	GA-ELM	14.471	0.827	1.099
	PSO-ELM	14.430	0.824	1.093
	GWO-ELM	14.419	0.823	1.091
	MVO-ELM	14.410	0.823	1.092
	IMVO-ELM	14.399	0.822	1.090
B	ELM	11.076	1.184	1.534
	GA-ELM	10.330	1.109	1.436
	PSO-ELM	10.309	1.107	1.434
	GWO-ELM	10.311	1.107	1.433
	MVO-ELM	10.292	1.105	1.433
	IMVO-ELM	10.285	1.104	1.430
C	ELM	13.329	0.700	0.941
	GA-ELM	12.399	0.661	0.903
	PSO-ELM	12.331	0.659	0.900
	GWO-ELM	12.324	0.658	0.902
	MVO-ELM	12.305	0.658	0.900
	IMVO-ELM	12.296	0.657	0.901
D	ELM	14.829	1.123	1.489
	GA-ELM	13.753	1.047	1.400
	PSO-ELM	13.706	1.042	1.395
	GWO-ELM	13.709	1.043	1.397
	MVO-ELM	13.686	1.042	1.396
	IMVO-ELM	13.665	1.040	1.394