Bayesian Inference For Semiparametric Regression Of Ordinal Response Data

A vast literature in Statistics,biometrics,and econometrics is concerned with the analysis of Ordinal response data,including the Bayesian inference for semiparametric binary regression(see Newton et al 1996,Mouchart and Scheihing 1998).Based on the early work, we propose a Bayesian regression model for ordinal response data that places no structural restrictions on the link function except being probability measure and not a mass point. Predictors enter linearly.We demonstrate Bayesian inference calculations in this model.We use Dirichlet process as a natural prior measure over this semiparametric model and P(?)lya sequence theory to formulate this measure in terms of a finite number of unobserved variables. We design a Gibbs Sampler algorithm for posterior simulation.In addition,we present the specific case when the response data do not have positive correlation with the latent variables.In this case,we will use the same model mentioned before to fit the samples after a transformation on the ordinal response data.