Short-Term Prediction of Air Pollution in Macau Using Support Vector Machines

Table 5

The tested results of SPM, SO₂, NO₂, and O₃ in different kernel models in winter experiment.

	Kernel
	Linear	Polynomial	RBF	Sigmoid	Wavelet
SPM
MAE	0.307	0.335	0.317	3.724	0.493
RMSE	0.397	0.489	0.425	10.392	0.606
CWIA	17.6%	20.5%	19.8%	90.3%	36.8%
RE	25.8%	29.0%	32.3%	45.2%	61.3%
SO2
MAE	0.823	1.024	0.811	32.112	1.126
RMSE	1.038	1.481	1.056	81.842	1.422
CWIA	17.5%	19.0%	17.9%	97.0%	33.0%
RE	32.3%	35.5%	35.5%	93.5%	51.6%
NO2
MAE	0.293	0.491	0.305	3.736	0.459
RMSE	0.391	0.807	0.375	9.950	0.624
CWIA	17.6%	26.0%	15.3%	102.2%	39.6%
RE	25.8%	41.9%	16.1%	71.0%	41.9%
O3
MAE	0.291	0.319	0.307	1.799	0.462
RMSE	0.369	0.388	0.369	2.217	0.571
CWIA	24.0%	27.7%	26.4%	76.1%	39.0%
RE	45.2%	51.6%	48.4%	93.5%	71.0%