| With the accelerating urbanization process,the construction and development of urban agglomeration has become an inevitable trend.In order to promote the sustainable development of urban agglomerations,attention should be paid to the important role played by the public transport system in the process of urban agglomerations construction and development.The development level of public transport system in urban agglomerations needs to be measured by accessibility evaluation index.Public transport development is systematic,and accessibility evaluation should be multi-faceted and multi-dimensional.Therefore,based on existing research and thinking,this paper sorts out the accessibility of public transport in urban agglomerations from different dimensions and selects evaluation indicators that can be measured in each dimension.Based on machine learning algorithm,an optimal model is constructed to explore the real relationship between public transport accessibility evaluation indicators and intercity travel mode selection behavior,so as to seek ways to improve public transport accessibility and sustainable transportation development in urban agglomerations.Based on “Shaanxi Provincial Natural Science Basic Research Program Project(2022JQ-735)” and “Ministry of Education Humanities and Social Science Research Youth Fund Project(19YJCZH024)”,this study first constructed an objective accessibility evaluation index for urban agglomeration public transport.It includes the accessibility evaluation index of urban public transport hub transfer,the accessibility evaluation index of urban agglomeration public transport based on intercity space-time barrier,based on transportation infrastructure level,based on operation service level,and the perceived accessibility evaluation index of urban agglomeration public transport.Secondly,based on the development status and travel characteristics of intercity traffic in Guanzhong Plain urban agglomerations,a questionnaire on intercity travel behavior of Guanzhong Plain urban agglomerations is designed and investigated.Based on the survey data,travel behavior information of urban agglomerations and evaluation data of perceived accessibility of public transportation in urban agglomerations are obtained.Based on survey data and other multi-source data platforms,the objective evaluation indicators of urban agglomeration public transport are measured.In this paper,random forest and XGBoost algorithms are used to construct a prediction model of residents’ intercity travel mode choice behavior.Combining the limitations and advantages of the two algorithms,a logistic regression based Stacking fusion model is further constructed.The prediction accuracy of the model was comprehensively compared and selected through the indexes of maximum accuracy of one time prediction,F1 value and 10-fold cross-validation results,and the nonlinear influence,threshold effect and synergistic effect of multi-dimensional public transport accessibility index on the travel mode selection behavior of urban agglomeration were explored by using the optimal model.It is found that the logistic regression based Stacking fusion model has good applicability.Most of the public transport accessibility evaluation indicators constructed in this paper have a significant nonlinear impact on whether travelers choose public transport mode during intercity travel,and can determine its threshold range.Most public transport accessibility combinations have complementary synergistic effects on intercity travel mode selection,but only the number of departures and the importance of public transport network city nodes have additional superimposed effects on mode selection.Based on the research results,improvement measures are proposed from the aspects of passenger hub transfer environment,intercity public transport operation frequency,intercity public transport network connectivity and service quality,which provides a foundation for the improvement of public transport accessibility in Guanzhong Plain urban agglomeration. |
PDF Full Text Request