Table 3 Predictive results of six deep learning models.

	PM_2.5			PM₁₀			O₃
	RMSE	MAE	R²	RMSE	MAE	R²	RMSE	MAE	R²
Fangshan monitoring station
LSTM	20.9	14.62	0.81	24.16	18.13	0.70	15.71	12.33	0.4
Seq2Seq	21.03	14.71	0.81	26.44	20.11	0.64	14.81	11.18	0.47
CNN-LSTM	25.74	18.4	0.72	29.44	23.00	0.55	17.06	13.43	0.3
GC-LSTM	22.06	15.95	0.79	24.56	18.97	0.69	17.82	14.48	0.33
SpAttRNN	22.62	15.58	0.78	24.41	19.24	0.69	16.4	12.48	0.43
AAMGCRN	19.11	13.58	0.84	23.73	18.07	0.71	14.74	11.07	0.54
Tiantan monitoring station
LSTM	20.64	13.46	0.81	23.87	17.58	0.66	16.91	12.92	0.35
Seq2Seq	21.49	13.35	0.79	22.35	16.52	0.71	15.67	11.71	0.44
CNN-LSTM	24.21	17.80	0.74	24.18	18.63	0.66	16.03	12.71	0.41
GC- LSTM	20.53	14.00	0.81	18.90	13.93	0.71	17.01	12.65	0.46
SpAttRNN	20.77	14.13	0.81	19.37	14.66	0.69	17.14	12.68	0.46
AAMGCRN	19.10	12.67	0.84	18.80	13.87	0.71	15.44	11.47	0.56
Dongsi monitoring station
LSTM	21.88	14.03	0.81	20.59	15.11	0.78	17.67	12.80	0.27
Seq2Seq	21.27	13.94	0.82	21.5	15.96	0.76	16.77	12.49	0.34
CNN-LSTM	24.46	17.38	0.77	24.65	19.11	0.69	20.83	18.57	-0.01
GC-LSTM	20.61	14.04	0.83	19.25	14.39	0.77	16.64	14.45	0.44
SpAttRNN	20.19	13.40	0.84	18.45	13.77	0.79	16.93	13.27	0.42
AAMGCRN	18.21	12.04	0.87	17.75	13.02	0.80	14.39	11.05	0.58

Quick links

Search