Development of Shopping Mode Choice Model with Bayesian Estimation Method

Abstract

In this paper, we developed shopping mode choice multinomial probit models basedon the shopping trip database in a suburban newtown with maximum likelihood estimation and Bayesian estimation using MCMC methods. According to our study, both estimation results were similar, while Bayesian estimation method could take account of standard deviation as well as mean of each coefficient. In addition, we found that the standard deviation for each explanatory parameter was smaller forthe multi-purpose shopping than the shopping-only case.