A New Hybrid Model Based on an Intelligent Optimization Algorithm and a Data Denoising Method to Make Wind Speed Predication

Table 6

Errors of different traditional models in February.

	Models	MAE (m/s)	MSE (m^2/s^2)	MAPE (%)	Running time (s)
Observation site 1	BPANN	0.3911	0.3041	6.68	2.8502
	ARIMA	0.4088	0.3208	7	47.2708
	AFSA-BPANN	0.3928	0.3069	6.71	105.3652
	WT-BPANN	0.3770	0.2735	6.58	5.1081
	WT-ARIMA	0.3819	0.2821	6.53	40.7528
	WAFSA-BPANN	0.3040	0.1622	5.4	156.3590
Observation site 2	BPANN	0.4219	0.3304	9.5	2.0659
	ARIMA	0.4415	0.3526	9.94	35.5574
	AFSA-BPANN	0.4307	0.3389	9.78	106.2827
	WT-BPANN	0.4087	0.3045	9	2.4668
	WT-ARIMA	0.4130	0.3109	9.03	42.0537
	WAFSA-BPANN	0.3134	0.1674	7.44	148.8582
Observation site 3	BPANN	0.3975	0.2845	8.27	0.5424
	ARIMA	0.4098	0.3021	8.43	42.4113
	AFSA-BPANN	0.3949	0.2793	8.24	105.7024
	WT-BPANN	0.3833	0.2654	8.05	0.7321
	WT-ARIMA	0.3849	0.2694	7.92	40.7567
	WAFSA-BPANN	0.3055	0.1601	6.87	146.8292