Spatio temporal random effect models for child labor mapping

Abstract

Count data are most commonly modeled using the Poisson model, or by one of its many extensions. In this study, a Poisson generalized linear mixed model (GLMM) with spatio-temporal random effects was modeled using two approaches. They are conditional autoregressive linear models and conditional autoregressive adaptive models. The models in this paper are fitted in a Bayesian setting using Markov chain Monte Carlo simulation. All parameters whose full conditional distributions have a closed-form distribution are Gibbs sampled, which includes the regression parameters and the random effects, as well as the variance parameters in all models. The models are applied to child labor in Sumatra between 2014 and 2017. Our main results show that the unemployment rate, illiteracy rates, and dropout rates influence the child labor number in Sumatra. Of the two models used, the spatio-temporal model with CAR adaptive is the best model by producing a simpler model but requires more time to build the model.