Table 1. Posterior model probabilities obtained using Gibbs Variable Selection (First stage).
Model | Environmental Variables | Binomial | ZIB |
1. | Night LST, NDVI | 2.46% | 2.52% |
2. | Night LST, NDVI, Area type | 72.21% | 74.28% |
3. | Night LST, Rainfall, NDVI, Area type | 12.13% | 13.23% |
4. | Others | 13.2% | 9.97% |
In the model with the highest posterior probability (72.21% with Binomial model and 74.28% with ZIB), Night LST, NDVI and Area type were included as covariates. This model was selected and used to predict the malaria risk.