Skip to main content
. 2017 Jun 1;16:232. doi: 10.1186/s12936-017-1874-0

Table 2.

Multivariate seasonal autoregressive integrated moving average (SARIMA) models of malaria incidence in four administrative areas in Swaziland

SARIMA modela Coefficients SE AIC AIC difference
Hhohho
 Malaria only 9.6
 Malaria + MEI (lag = 2) −0.067 0.049 11.67 2.07
 Malaria + TMAX (lag = 0) 0.0137 0.0152 10.97 1.37
 Malaria + TMIN (lag = 3) 0.0124 0.0089 12.58 2.98
 Malaria + precipitation (lag = 3) 0.02 0.0066 5.43 b 4.17
Lubombo
 Malaria only 92.34
 Malaria + MEI (lag = 1) −0.2039 0.05 89.46 −2.88
 Malaria + TMAX (lag = 3) 0.0449 0.0775 88.88 −3.46
 Malaria + TMIN (lag = 1) 0.0135 0.0092 92.22 −0.12
 Malaria + precipitation (lag = 2) 0.0224 0.0007 86.59 b 5.75
Manzini
 Malaria only 474.99 b
 Malaria + MEI (lag = 3) 0.0054 0.0085 −471.65 3.34
 Malaria + TMAX (lag = 3) 0.7475 0.3119 −471.83 3.16
 Malaria + TMIN (lag = 2) 0.0004 0.0024 −471.27 3.72
 Malaria + precipitation (lag = 1) 0.0054 0.0025 −471.12 3.87
Shiselweni
 Malaria only 396.8 b
 Malaria + MEI (lag = 7) 0.0156 0.0139 −375.52 21.28
 Malaria + TMAX (lag = 4) 0.009 0.004 −389.54 7.26
 Malaria + TMIN (lag = 2) 0.0039 0.0023 −393.72 3.08
 Malaria + precipitation (lag = 3) 0.006 0.0032 −390.88 5.92

aThe lag is selected using a cross-correlation function

bThe model with the lowest AIC value is indicated in italic type