In many situations discrete count data have a large proportion of zero more than the general model proportion of zero. In that case we use the zero-inflated model. Zero-inflated model is assume that a count response variable is assumed to be distribut...
In many situations discrete count data have a large proportion of zero more than the general model proportion of zero. In that case we use the zero-inflated model. Zero-inflated model is assume that a count response variable is assumed to be distributed as a mixture of a distribution with point mass of one at zero and a Poisson distribution or Negative binomial distribution. In this paper, we propose a Bayesian inference for the zero-inflated Poisson(ZIP) regression model and zero-inflated Negative binomial(ZINB) regression model by using a Markov Chain Monte Carlo methods, specially we use Metropolis Hastings algorithm and Gibbs sampling. We applied the real crash count data and conduct the model selection using Bayes factors. We choose the ZIP regression model for this data. And we also conduct parameter estimation for the ZIP regression model.