Short-Term Wind Speed Prediction Using EEMD-LSSVM Model

Table 3

Performance comparison of different models for forecasting wind speed (Case One).

Prediction horizon	Indicators	LSSVM	ARIMA	EMD-LSSVM	BP	EEMD-ARIMA	EEMD-LSSVM
One-step	RMSE	0.7981	0.8006	0.4525	0.8206	0.4015	0.3819
	MAPE	10.08%	10.28%	5.066%	10.58%	4.985%	4.435%
	MAE	0.5847	0.5914	0.3038	0.6126	0.2889	0.2634
	CC	0.9430	0.9426	0.9822	0.9399	0.9859	0.9872
	SSE	91.7192	92.2980	29.4868	96.9681	23.2115	21.0020
	SDE	0.7979	0.7992	0.4515	0.8204	0.4009	0.3830
Two-step	RMSE	1.1612	1.1669	0.5566	1.2059	0.5284	0.4923
	MAPE	14.58%	15.37%	6.53%	16.59%	6.80%	5.90%
	MAE	0.8536	0.8752	0.3877	0.9405	0.4010	0.3418
	CC	0.8765	0.8741	0.9732	0.8662	0.9753	0.9788
	SSE	192.8291	194.7302	44.3054	207.9640	39.9335	34.6568
	SDE	1.1605	1.1632	0.5586	1.2049	0.5313	0.4884
Three-step	RMSE	1.4282	1.4341	0.7690	1.4586	0.7153	0.6836
	MAPE	18.08%	19.38%	9.557%	20.57%	9.677%	7.954%
	MAE	1.0342	1.0679	0.5667	1.1439	0.5496	0.4582
	CC	0.8075	0.8021	0.9473	0.7986	0.9576	0.9583
	SSE	289.6285	292.0333	83.9772	302.0969	72.6622	66.3587
	SDE	1.4258	1.4246	0.7786	1.4555	0.7177	0.6724