Table 5.

Averaged metrics calculated on the test sets during cross-validation 2019.

Algorithm	CO		NO2		PM10
	${\mathit{R}}^2$	RMSE	${\mathit{R}}^2$	RMSE	${\mathit{R}}^2$	RMSE
Linear regression	0.935	0.066	0.737	13.412	0.837	12.551
Neural network 1 (2 HL 1)	0.941	0.065	0.869	9.450	0.839	12.583
Neural network 2 (3 HL)	0.943	0.063	0.872	9.344	0.850	12.124
AdaBoost	0.924	0.074	0.843	10.360	0.846	14.560
Random forest	0.945	0.060	0.894	8.540	0.872	11.123
SVM	0.933	0.070	NC 2	NC	0.835	12.748

¹ HL, hidden layer; ² NC, non-convergent.