Table 2.

Alternative ARIMAX models for the three cities

City	Model	Normalized BIC value	P*	MAPE (%)a
Xuzhou	ARIMA (1,1,1)(0,1,1)12	8.857	0.861	12.54
	ARIMA (1,1,1)(0,1,1)12 + MAS2	8.595	0.714	14.05
	ARIMA (1,1,1)(0,1,1)12 + MAH1	8.467	0.399	24.09
	ARIMA (1,1,1)(0,1,1)12 + MP2	8.617	0.356	11.96
	ARIMA (1,1,1)(0,1,1)12 + MST1	8.593	0.767	17.62
	ARIMA (1,1,1)(0,1,1)12 + MAS2 + MAH1	8.609	0.338	25.02
	ARIMA (1,1,1)(0,1,1)12 + MAS2 + MP2	8.658	0.691	17.22
	ARIMA (1,1,1)(0,1,1)12 + MAS2 + MST1	8.679	0.902	17.34
	ARIMA (1,1,1)(0,1,1)12 + MAH1 + MP2	8.560	0.431	20.68
	ARIMA (1,1,1)(0,1,1)12 + MAH1 + MST1	8.604	0.416	24.30
	ARIMA (1,1,1)(0,1,1)12 + MP2 + MST1	8.674	0.751	17.55
	ARIMA (1,1,1)(0,1,1)12 + MAS2 + MAH1 + MP2	8.700	0.371	20.71
	ARIMA (1,1,1)(0,1,1)12 + MAS2 + MAH1 + MST1	8.755	0.427	23.01
	ARIMA (1,1,1)(0,1,1)12 + MAS2 + MP2 + MST1	8.755	0.851	17.21
	ARIMA (1,1,1)(0,1,1)12 + MAH1 + MP2 + MST1	8.692	0.241	39.17
	ARIMA (1,1,1)(0,1,1)12 + MAS2 + MAH1 + MP2 + MST1	8.831	0.581	17.44
Nantong	ARIMA (0,1,1)(0,1,1)12	8.609	0.433	15.57
	ARIMA (0,1,1)(0,1,1)12 + MAT0	8.288	0.981	16.77
	ARIMA (0,1,1)(0,1,1)12 + MAP1	8.183	0.777	11.16
	ARIMA (0,1,1)(0,1,1)12 + MAS2	8.323	0.730	16.29
	ARIMA (0,1,1)(0,1,1)12 + MAT0 + MAP1	8.340	0.836	14.99
	ARIMA (0,1,1)(0,1,1)12 + MAT0 + MAS2	8.419	0.965	16.97
	ARIMA (0,1,1)(0,1,1)12 + MAP1 + MAS2	8.314	0.766	11.90
	ARIMA (0,1,1)(0,1,1)12 + MAT0 + MAP1 + MAS2	8.470	0.892	13.06
Wuxi	ARIMA (0,1,1)(0,1,1)12	6.933	0.176	9.70
	ARIMA (0,1,1)(0,1,1)12 + MAH0	6.845	0.119	9.66
	ARIMA (0,1,1)(0,1,1)12 + MST0	6.818	0.068	10.51
	ARIMA (0,1,1)(0,1,1)12 + MAH0 + MST0	7.003	0.088	9.74

BIC Bayesian information criterion, MAPE mean absolute percentage error, MAT monthly average temperature; MAP monthly average atmospheric pressure, MAS monthly average wind speed, MAH monthly average relative humidity, MP Monthly precipitation, MST monthly sunshine time, 0 0-month lag, 1 1-month lag, 2 2-month lag

^*Ljung-Box test

^a MAPE of the model in predicting the monthly number of PTB cases in 2018