Research on PM2.5 Spatiotemporal Forecasting Model Based on LSTM Neural Network

Table 3

Comparison of various baseline models with the model’s indicators for the next hour.

Baseline model	MAE (μg/m^3)	RMES (μg/m^3)	R^2
LR	13.08	23.707	0.903
RF	8.02	15.131	0.918
SVM	34.69	68.106	0.255
ARMA	7.46	13.958	0.923
LSTM_N	7.11	13.943	0.922
LSTM_NW	7.33	13.567	0.926
LSTM_NE	7.79	12.757	0.919
LSTM_WE	7.21	12.353	0.926
LSTM_S	7.81	14.664	0.916
ST_LSTM	6.81	12.080	0.930