Table 2 Results of the INLA-BYM zero-inflated negative binomial model with ridge regression for triatomine abundance. Estimates of fixed effects (Population and deforested area) are presented as posterior mean and 95% credible interval (0.025quant – 0.975quant). The random effects include independent and identically distributed (IID) priors for forest area to account for collinearity through ridge regression, as well as a Besag-York-Mollié (BYM) model for Spatial effects. The model hyperparameters include precision estimates for random effects and the zero-inflation component of the negative binomial distribution.
From: Deforestation effects and house invasion by chagas disease vectors in Brazil
Fixed effects | mean | 0.025quant | 0.975quant |
|---|---|---|---|
(Intercept) | -1.50 | -2.91 | -0.22 |
Population (1000 hab) | -0.06 | -0.11 | -0.01 |
Deforested area (Km2) | |||
Model hyperparameters | |||
size for nbinomial_1 zero-inflated observations | 1265.09 | 2.87 | 7944.53 |
zero-probability parameter for zero-inflated nbinomial_1 | 0.173 | 0.038 | 0.421 |
Precision for the Forest area | 0.219 | 0.122 | 0.359 |
Precision of the IID component | 2219.266 | 158.71 | 8541.932 |
Precision of spatial component | 2246.76 | 153.499 | 8825.007 |
