Table 2.

Autocorrelation and partial correlation of monthly malaria incidence, and the fitting and predictive residuals of the model

Lag	Monthly malaria incidence					Fitting residual					Predictive residual
	AC	PAC	LB	P		AC	PAC	LB	P		AC	PAC	LB	P
1	0.84	0.84	102.48	< 0.01		-0.06	-0.06	0.20	0.66		0.29	0.29	1.25	0.26
2	0.60	-0.33	155.20	< 0.01		-0.04	-0.05	0.31	0.86		-0.33	-0.45	3.12	0.21
3	0.30	-0.33	168.51	< 0.01		0.06	0.06	0.54	0.91		-0.13	0.18	3.45	0.33
4	0.03	-0.07	168.64	< 0.01		0.03	0.03	0.59	0.97		0.17	0.01	4.04	0.40
5	-0.19	-0.05	173.85	< 0.01		0.04	0.05	0.67	0.99		0.07	-0.03	4.17	0.53
6	-0.27	0.17	185.02	< 0.01		0.01	0.02	0.68	0.99		-0.02	0.09	4.18	0.65
7	-0.24	0.12	194.14	< 0.01		0.05	0.06	0.88	0.99		-0.17	-0.26	5.11	0.65
8	-0.11	0.15	196.02	< 0.01		-0.04	-0.04	0.98	0.99		-0.28	-0.18	8.47	0.39
9	0.09	0.18	197.33	< 0.01		0.07	0.07	1.32	0.99		-0.19	-0.16	10.47	0.31
10	0.32	0.20	213.76	< 0.01		-0.14	-0.14	2.62	0.99		0.06	-0.02	10.80	0.37
11	0.51	0.08	254.12	< 0.01		0.06	0.06	2.92	0.99
12	0.56	-0.18	303.34	< 0.01		-0.22	-0.25	6.51	0.89

AC: autocorrelation coefficient. PAC: partial autocorrelation coefficient. LB: Ljung-Box Q Statistic. Lag: the number of lagged months. For the monthly malaria incidence, P < 0.05 indicates a strong autocorrelation of monthly malaria incidence. For the fitting and predictive residuals, P > 0.05 indicates that the model extracted the information sufficiently and had good prediction validity.