Predictive Study of Tuberculosis Incidence by ARMA Model Combined with Air Pollution Variables

Table 3

Parameter estimates of the tentative models with their AIC and SC.

Model		Variable	Coefficient	Std. error	t-statistic		AIC	SC
ARMA (1, 1)		C	5.41	0.46	11.64	<0.01	2.57	2.69
		AR (1)	0.79	0.12	6.51	<0.01
		MA (1)	−0.23	0.20	−1.14	0.26
ARMA (1, 2)		C	5.35	0.45	11.82	<0.01	2.56	2.71
		AR (1)	0.72	0.16	4.53	<0.01
		MA (1)	−0.21	0.20	−1.05	0.29
		MA (2)	0.25	0.17	1.45	0.16
ARMA (1, (2))		C	5.32	0.39	13.69	<0.01	2.53	2.65
		AR (1)	0.602	0.13	4.69	<0.01
		MA (2)	0.28	0.15	1.82	0.08
ARMA (1, 3)		C	5.37	0.46	11.68	<0.01	2.60	2.80
		AR (1)	0.74	0.216	3.48	<0.01
		MA (1)	−0.214	0.266	−0.81	0.42
		MA (2)	0.234	0.186	1.291	0.20
		MA (3)	−0.06	0.21	−0.27	0.79
ARMA (1, (3))		C	5.72	0.15	37.39	<0.01	2.46	2.58
		AR (1)	0.802	0.07	11.62	<0.01
		MA (3)	−0.93	0.046	−20.18	<0.01
ARMA (1, (1, 3))		C	5.79	0.15	38.68	<0.01	2.39	2.55
		AR (1)	0.86	0.05	17.20	<0.01
		MA (1)	−0.48	0.11	−4.38	<0.01
		MA (3)	−0.48	0.11	−4.41	<0.01
ARMA (1, (2, 3))		C	5.36	0.39	13.50	<0.01	2.57	2.73
		AR (1)	0.60	0.13	4.45	<0.01
		MA (2)	0.28	0.16	1.80	0.08
		MA (3)	0.01	0.15	0.08	0.92