A Multiple Kernel Learning Approach for Air Quality Prediction

Table 15

Performance comparison for predicting the future 9 hour’s PM2.5 IAQL in Beijing.

	Accuracy	mse	wr	wf	wp
ARIMA	0.410	1.208	0.471	0.472	0.474
RF	0.457	1.909	0.457	0.446	0.439
MLP	0.48	1.706	0.48	0.457	0.456
SVC_linear	0.45	2.236	0.45	0.385	0.424
SVC_rbf	0.492	1.746	0.492	0.452	0.456
SVC_poly	0.482	1.813	0.482	0.453	0.45
SVC_sigmoid	0.39	4.000	0.39	0.219	0.152
LSTM	0.390	4.000	0.391	0.217	0.151
MKSVC	0.507	1.133	0.510	0.505	0.500