Determining route-choice behaviour of public transport passengers using Bayesian statistical inference

Using smart card systems for public transport fare collection has provided a great opportunity to access large-scale and high-quality travel data of transit users. This data has the potential to be used for modelling passenger behaviour. In this paper, Bayesian statistical inference is used to model passenger route-choice behaviour and to estimate attributes of travel-time components. The Bayesian approach provides a comprehensive posterior knowledge of the system. The posterior density integrates the observed passenger travel data with our prior knowledge about the transit network. Due to the high dimensional nature of the parameter space, the Markov Chain Monte Carlo Method is utilised to compute the mean value for each parameter. The suggested model is calibrated and validated using the fare card data from Brisbane, Australia, for March 2013. The dataset contains those journeys with an educational purpose. The inference results for link travel times are compared with the scheduled times from Brisbane's General Transit Feed Specification to show the reliability of public transport services. The estimated route-choice parameters indicate that transfer time is seven times more important than in-vehicle time for transit passengers