A Multiple Kernel Learning Approach for Air Quality Prediction

Table 10

Performance comparison for predicting the future 9 hour’s AQHI in HK.

	Accuracy	mse	wr	wf	wp
ARIMA	0.467	2.020	0.467	0.433	0.436
RF	0.738	0.332	0.738	0.735	0.737
MLP	0.777	0.255	0.777	0.776	0.776
SVC_linear	0.799	0.231	0.799	0.798	0.799
SVC_rbf	0.787	0.244	0.787	0.786	0.786
SVC_poly	0.785	0.25	0.785	0.784	0.784
SVC_sigmoid	0.267	4.991	0.267	0.113	0.071
LSTM	0.681	0.393	0.692	0.676	0.645
MKSVC	0.817	0.203	0.821	0.815	0.820