A Machine Learning Approach to Predict Air Quality in California

Table 5

Random search optimal parameter results per pollutant dataset.

Pollutants	PCA dataset				Normalized dataset
	Kernel	Opt C	Opt ε	R ^2	Kernel	Opt C	Opt ε	R ^2
CO	RBF	3	0.08	0.783	RBF	2	0.033	0.916
NO2	RBF	3	0.025	0.882	RBF	1	0.067	0.948
SO2	RBF	1	0.062	0.712	RBF	2	0.086	0.718
Ozone	RBF	2	0.076	0.903	RBF	2	0.02	0.979
PM2.5	RBF	1	0.055	0.765	RBF	3	0.032	0.767