ABSTRACT
Bayesian models, which integrate of Bayes’ theorem with classical statistical models, have become increasingly popular in the transportation field. This chapter gives an introduction to Bayesian by first discussing Bayes’ theorem and its potential application to transportation data analysis. The chapter then goes on to give an extensive discussion of Markov Chain Monte Carlo (MCMC) based model estimation. MCMC is a sampling-based approach to estimation that is well suited for Bayesian models and enables the estimation of complex functional forms which can be sometimes difficult to estimate using maximum likelihood methods. Convergence and identifiability issues are discussed in detail along with goodness-of-fit measures, sensitivity analysis, and model section criterion. Numerous examples are provided to support these discussions. This chapter clearly shows the potential of Bayesian approaches as a viable modeling alternative to traditional statistical methods.
