Skip to main content
. 2018 Aug 13;13(8):e0201700. doi: 10.1371/journal.pone.0201700

Table 2. Summary of models fitted to measles data for Namibia and their corresponding deviance information criterion (DIC) and sum of log of conditional predictive ordinate (SumLog(CPO)).

ICAR—intrinsic conditional autoregressive structure model, CAR—Conditional Autoregressive, IID—identically distributed.

Poisson model Spatial component Fixed effect D p DIC SumLog(CPO)
1 ICAR None 14617.09 94.73 14806.55 -7558.995
2 CAR None 14616.81 94.96 14806.73 -7559.032
3 IID None 14613.48 99.71 14812.9 -7577.505
4 ICAR All covariates (parametric model) 13173.53 152.61 13478.75 -7104.421
5 CAR All covariates (parametric model) 13173.23 152.76 13478.75 -7104.024
6 IID All covariates (parametric model) 13170.76 166.18 13503.12 -7166.587
7 ICAR All covariates (non-parametric model) 7532.17 104.02 7740.21 -4297.502
8 CAR All covariates (non-parametric model) 7531.79 104.20 7740.19 -4298.066
9 CAR All covariates +time-space interaction (Type I) 3765.80 617.93 5001.66 -2853.179
10 ICAR All covariates + time-space interaction (Type I) 3765.58 618.01 5001.60 -2852.438
11 CAR All covariates + time-space interaction (Type I) + measurement error models 3765.35 617.16 4999.69 -2850.238
12 ICAR All covariates + time-space interaction (Type I) + measurement error model 3748.53 629.33 5007.19 -2893.197
13 CAR Covariates + time-space interaction (Type II) + measurement error models 3766.41 617.16 5000.73 -2835.843
14 ICAR Covariates + time-space interaction (Type II) + measurement error models 3782.19 608.52 4999.23 -2831.314