A Multiple Kernel Learning Approach for Air Quality Prediction

Table 14

Performance comparison for predicting the future 6 hour’s PM2.5 IAQL in Beijing.

	Accuracy	mse	wr	wf	wp
ARIMA	0.442	2.381	0.442	0.367	0.332
RF	0.468	2.024	0.468	0.433	0.437
MLP	0.493	1.67	0.493	0.463	0.462
SVC_linear	0.451	2.207	0.451	0.385	0.408
SVC_rbf	0.500	1.595	0.500	0.477	0.478
SVC_poly	0.490	1.844	0.490	0.435	0.441
SVC_sigmoid	0.391	3.999	0.391	0.219	0.153
LSTM	0.397	3.701	0.396	0.253	0.167
MKSVC	0.513	1.275	0.513	0.520	0.519