Short-Term Wind Speed Prediction Using EEMD-LSSVM Model

Table 4

The percentage improvements of EEMD-LSSVM over other models (Case One).

Prediction horizon	Indicators (improvement)	Compared models
		LSSVM	ARIMA	BP	EMD-LSSVM	EEMD-ARIMA
One-step	RMSE	52.15%	52.30%	53.46%	15.60%	4.88%
	MAPE	56.00%	56.86%	58.08%	12.46%	11.03%
	MAE	54.95%	55.46%	57.00%	13.30%	8.83%
	CC	4.69%	4.73%	5.03%	0.51%	0.13%
	SSE	77.10%	77.25%	78.34%	28.77%	9.52%
	SDE	52.00%	52.08%	53.32%	15.17%	4.46%
Two-step	RMSE	57.60%	57.81%	59.18%	11.55%	6.83%
	MAPE	59.53%	61.61%	64.44%	9.65%	13.24%
	MAE	59.96%	60.95%	63.66%	11.84%	14.76%
	CC	11.67%	11.98%	13.00%	0.58%	0.36%
	SSE	82.03%	82.20%	83.34%	21.78%	13.21%
	SDE	57.91%	58.01%	59.47%	12.57%	8.07%
Three-step	RMSE	52.14%	52.33%	53.13%	11.11%	4.43%
	MAPE	56.01%	58.96%	61.33%	16.77%	17.81%
	MAE	55.70%	57.09%	59.94%	19.15%	16.63%
	CC	18.67%	19.47%	20.00%	1.16%	0.07%
	SSE	77.09%	77.28%	78.03%	20.98%	8.68%
	SDE	52.84%	52.80%	53.80%	13.64%	6.31%