Metro Transfer Perception Time Research On The Basis Of Bayesian Network Modelling Analysis

With the rapid expansion of the metro network in modern urban areas,transfer experience of metro passengers is often poor due to improper transfer design and sudden increase in transfer passenger flow in actual operations.Therefore,how to improve the transfer experience of metro passengers and the transfer service level of the metro system through reasonable and effective operation management have become an urgent problem for operators to pay attention to and solve.From the perspective of metro passengers,this article analyses and explores the factors that affect passengers’ transfer experience,and then provides theoretical analysis basis and practical operation reference for metro operators to develop transfer service improvement plans meeting the needs of passengers.In view of the intuitive influence of seasonal factors on the travel experience of public transport passengers,this research mainly explores the strategies to improve the transfer experience of metro passengers in different seasons by modelling analysis of the metro transfer perception time in different seasons.Based on the questionnaire survey combined with the field survey,the transfer perceived and actual data of metro passengers are collected to firstly analyse the difference in transfer perception of metro passengers in different seasons.Then,this study uses Bayesian network(BN)method to model and analyse the metro transfer perception time in different seasons,and proposes an improved PC algorithm to learn the structure of BN.After BN learning,the classification prediction evaluation indicators are introduced to evaluate the predictive ability of the constructed model,and the concept of effective evidence and variable function is proposed to verify its validity.After that,this work quantitatively explores the effects of factors affecting the metro transfer perception time in different seasons through the sensitivity analysis.Finally,the strategies for reducing the metro transfer perception time in different seasons are proposed.The results of the study show that passengers in spring and autumn have a more serious perceived overestimation of metro transfer activities than passengers in summer and winter.In addition,models of metro transfer perception time in different seasons constructed in this work can predict the transfer perception time of passengers well,and then verifies the performance of the effectiveness of the proposed improved PC algorithm from the side.Finally,the comparisons of structures and sensitivity analysis results of BN models in different seasons show that there are certain differences in the influencing factors and their effects of metro transfer perception time in different seasons.Therefore,in order to reduce the transfer perception time of passengers in spring,the most effective strategy should be to reduce transfer waiting time and improve the transfer environment.In summer,since passengers pay more attention to the transfer walking process,if the transfer experience of passengers is improved,alleviating the congestion of transfer walking process and reducing the transfer walking time should be given consideration.In addition,the combination of alleviating the congestion of the transfer walking process,improving the transfer environment,and reducing the walking and waiting time of the transfer has the best effect in reducing the transfer perception time of passengers in the autumn.Finally,as for the transfer experience of winter passengers,we should focus on improving the crowdedness and time-consuming of transfer walking and waiting process.This paper contains 28 figures,21 tables and 63 references.