The Integrated Nested Laplace Approximation for Fitting Dirichlet Regression Models

This article introduces a Laplace approximation to Bayesian inference in Dirichlet regression models, which can be used analyze set of variables on simplex exhibiting skewness and heteroscedasticity, without having transform the data. These data, mainly consist proportions or percentages disjoint categories, are widely known as compositional data common areas such ecology, geology, psychology. We provide both theoretical foundations description how implemented case regression. The also package dirinla R-language that extends R-INLA package, not deal directly with likelihoods. Simulation studies presented validate good behavior proposed method, while real case-study is show this approach applied. Supplementary materials for available online.