Modeling Mode Choice Behaviors for Public Transport Commuters in Beijing

Abstract

Based on an analysis of the factors that influence the mode choice behaviors of commuters who use public transport (subways and buses), this study developed a questionnaire using a combination of revealed preference (RP) and stated preference (SP) techniques; both online and paper surveys were conducted to gather commuters’ travel choices between subways and buses. A binary logit (BL) specification was proposed to examine public transport commuters’ travel choices. The regression coefficients were estimated using maximum likelihood estimation. Finally, based on the data obtained from Beijing public transport smart cards, a support vector machine (SVM) classification model was established to identify commuters’ mode choices, and the accuracy was found to be as high as 94.24%. The estimated mode choice model was employed to predict the market shares of both subways and buses after a new fare scheme was implemented. The results showed that the model had high prediction accuracy: the average absolute error for predicting the market share of buses was 5.93%.