A New Hybrid Model FPA-SVM Considering Cointegration for Particular Matter Concentration Forecasting: A Case Study of Kunming and Yuxi, China

Table 1

Statistical parameters of PM_2.5 and PM₁₀ in each data set.

Particular matter	Region	Min	Max	Mean	Std	SK	CV
PM2.5	Kunming
	All	7.70	95.80	28.34	12.33	0.98	0.43
	Training	7.70	95.80	30.45	12.26	0.95	0.40
	Test	7.80	35.50	18.95	7.20	0.55	0.38
	Yuxi
	All	4.80	91.00	24.20	12.06	1.23	0.50
	Training	5.90	91.00	26.28	12.00	1.25	0.46
	Test	4.80	36.00	14.92	6.91	0.79	0.46
PM10	Kunming
	All	14.00	129.00	53.12	20.72	0.69	0.39
	Training	14.00	129.00	56.16	20.66	0.63	0.37
	Test	15.00	82.50	39.55	14.74	0.89	0.37
	Yuxi
	All	6.80	121.00	42.18	17.85	0.52	0.42
	Training	6.80	121.00	45.19	17.27	0.51	0.38
	Test	10.30	64.80	28.76	13.81	0.85	0.48

Min: the minimum. Max: the maximum. Std: the standard deviation. SK: the skewness. CV: the coefficient of variation.