Improve the Bayesian generalized latent variable models with non-linear variable and covariate of dichotomous data

Abstract

In this paper, we develop generalized latent variable models with non-linear variable and covariate. Dichotomous variables and covariates are used in this research, and the Gibbs sampling method (Markov chain Monte-Carlo simulation) is applied for estimation. The deviance information Criterion (DIC) is used as a model comparison statistics. Truncated normal distribution is used to handle the problem of dichotomous data in variables and covariates. Statistical analyses, which include the estimation of parameters, standard deviations and their highest posterior density, are explained. The proposed method is discussed using dichotomous data with the findings extracted from the OpenBUGS software