The Ordered Logistic Model With Dirichlet Priors

Bayesian theory is widely used in many areas, and it has been an important part of statistics. In this paper, we apply nonparametric method and Bayesian theory to ordered logistic model. We use nonparametric dirichlet process to measure the heter-ogeneity between different individuals, apply nonparametric Bayesian methods to ordered logistic model in order to improve ordered probit model and sample with MCMC methods from posterior distribution. Finally, we test the nonparametric Bayesian ordered logistic model with simulation data. From the difference between different models, we find that the ordered logistic model with dirichlet priors can measure the heterogeneity and it is better than ordered probit model in solving practical problems.