Evaluating the Service Quality of Public Transit from Passengers' Perceptions Using Bayesian Networks

Jingxian Wu,Min Yang,Zhibin Li,Cheng Ling,Xianyao Hou
2014-01-01
Abstract:Existing studies indicated that attributes of service qualities affect passengers’ perceptions on public transit. But most of previous models ignored the interrelationships among service attribute variables which could have a synergy on the overall satisfaction on public transit. Therefore, this paper aims to apply a new approach by Bayesian networks (BN) to capture the potential correlations among contributing factors in the evaluation of the service quality of public transit from passengers’ perceptions. Based on the data obtained from the 2013 Regular Bus service Survey in Nanjing, China, a BN was developed with combination of network structure and conditional probability table. Sensitivity analysis was conducted to identify the effects of each contributing factor in the BN. The results showed that safety attributes are less important to the passengers’ perceived satisfaction. Aspects in different categories are related tightly and there is a synergy which exerts a combined influence on the overall satisfaction. Currently five service aspects that most need to improve are “punctuality”, “accurate arriving information of buses”, “seats available”, “cleanliness onboard” and “convenience of transfers”. The experimental results showed that the method was effective and the analysis results can be assistances for policy makers.
What problem does this paper attempt to address?