Table 4.

The statistical evaluation of different model performances (PM2.5).

City	Model	RMSE	MAE	R2
Binzhou	SVM	21.25	19.86	0.66
	CEEMDAN-SVM	18.65	17.11	0.73
	BP	19.36	12.65	0.61
	MLP	22.95	15.56	0.58
	PSO-LSTM	26.94	18.97	0.45
	PSO-CNN-LSTM	26.32	18.68	0.47
	CEEMDAN-PSO-LSTM	17.69	12.89	0.76
	CEEMDAN-PSO-CNNLSTM	12.68	9.60	0.86
Jinan	SVM	23.13	21.34	0.52
	CEEMDAN-SVM	21.52	19.90	0.59
	BP	19.90	13.03	0.58
	MLP	30.00	20.84	0.46
	PSO-LSTM	24.28	16.48	0.47
	PSO-CNN-LSTM	23.94	16.47	0.49
	CEEMDAN-PSO-LSTM	16.04	11.21	0.77
	CEEMDAN-PSO-CNNLSTM	11.01	8.41	0.87
Handan	SVM	23.52	21.17	0.67
	CEEMDAN-SVM	16.14	13.20	0.86
	BP	26.57	17.69	0.39
	MLP	27.14	18.93	0.38
	PSO-LSTM	31.48	21.49	0.46
	PSO-CNN-LSTM	31.16	21.72	0.47
	CEEMDAN-PSO-LSTM	22.31	15.10	0.72
	CEEMDAN-PSO-CNNLSTM	12.94	9.99	0.88
Taiyuan	SVM	23.55	21.91	0.62
	CEEMDAN-SVM	19.86	17.92	0.73
	BP	30.55	16.26	0.25
	MLP	27.47	19.24	0.37
	PSO-LSTM	30.79	20.63	0.44
	PSO-CNN-LSTM	27.31	18.81	0.48
	CEEMDAN-PSO-LSTM	20.79	15.16	0.70
	CEEMDAN-PSO-CNNLSTM	12.38	9.33	0.88
Xinxiang	SVM	23.68	21.43	0.53
	CEEMDAN-SVM	18.53	14.85	0.71
	BP	725.29	19.29	0.34
	MLP	28.44	20.65	0.36
	PSO-LSTM	23.91	17.52	0.52
	PSO-CNN-LSTM	23.29	18.09	0.55
	CEEMDAN-PSO-LSTM	16.00	11.53	0.79
	CEEMDAN-PSO-CNNLSTM	11.63	8.98	0.88
Zibo	SVM	19.95	18.36	0.70
	CEEMDAN-SVM	19.70	17.84	0.71
	BP	23.49	16.96	0.48
	MLP	24.57	18.16	0.43
	PSO-LSTM	26.89	18.26	0.46
	PSO-CNN-LSTM	24.61	17.00	0.55
	CEEMDAN-PSO-LSTM	19.77	14.29	0.71
	CEEMDAN-PSO-CNNLSTM	10.66	8.34	0.89