MCMC Bayes-Mixed Logit for corridor transport mode split

Market share is a key indicator of competitiveness especially for a new band or mode. In order to develop effective marketing strategies for high-speed railway operation, the discrete choice concept model for mode split is constructed with seven elements based on the application of consumer choice theory. The MCMC Bayes-Mixed Logit algorithm is proposed analyzing a designed travel survey with intercity travel behavior information taking occupation and income into the utility function and applying the MCMC method with Matlab language. Finally, take the Beijing-Shanghai direct passenger traffic as an example with the real survey data. The experiment result shows that the fitting accuracy of Bayes-Mixed Logit is improved by 18.2% comparing with the MLE-Logit model.