A Novel Phase Space Reconstruction- (PSR-) Based Predictive Algorithm to Forecast Atmospheric Particulate Matter Concentration

Table 1

Performance of the predicted models in terms of RMSE and MAE.

Models	PM2.5 (μgm^−3)				PM10.0 (μgm^−3)				PM2.5/PM10.0
	Using proposed setting of the target set		Using traditional setting of the target set		Using proposed setting of the target set		Using traditional setting of the target set		Using proposed setting of the target set		Using traditional setting of the target set
	RMSE	MAE	RMSE	MAE	RMSE	MAE	RMSE	MAE	RMSE	MAE	RMSE	MAE
RF	1.27	0.86	8.82	5.07	1.60	1.16	19.86	9.80	0.02	0.02	0.21	0.16
SVR-L	0.10	0.09	12.65	9.93	0.06	0.06	141.5	140.0	0.05	0.04	0.71	0.59
SVR-R	24.04	21.08	21.85	18.43	37.80	34.17	62.17	60.39	10.14	9.74	5.58	5.02
FFNN	0.04	0.04	8.03	4.46	0.13	0.09	18.05	8.57	0.001	0.001	0.20	0.16