Research on influencing factors of transportation mode selection behavior in urban agglomerations considering travel preferences

WANG Jianjun; LI Jingtao; ZHANG Yuhui; LONG Xueqin

doi:10.11860/j.issn.1673-0291.20230140

Abstract：This study investigates the factors influencing passenger mode choice behavior in urban ag-glomerations,addressing the unclear mechanisms of such behavior. A Revealed Preference (RP) sur-vey is conducted on travel behavior within the Guanzhong urban agglomeration using a questionnaire,and the data's validity and reliability are tested. A travel mode selection prediction model considering travel preferences is constructed using the Gradient Boosting Decision Tree (GBDT) model,based on the survey data. The interpretable machine learning framework,SHapley Additive exPlanations (SHAP),is employed to analyze the prediction results. The constructed model considers actual passen-ger decision-making logic,deconstructing multiple travel mode choices into multiple binary choices to enhance model interpretability. The results indicate that the GBDT prediction model considering travel preferences effectively predicts travel mode choice behavior in urban agglomerations. Passenger prefer-ences for convenience,safety,speed and comfort,as well as car ownership home location and travel distance,significantly influence the binary choice behavior between private cars and public transporta-tion. The choices between high-speed buses and railway systems,as well as between ordinary rail-ways and high-speed trains,are significantly influenced by travel distance and passenger preferences for comfort,safety,speed,and punctuality. Car ownership and household location significantly affect the choice between high-speed buses and railway systems,while income status significantly influences the choice between ordinary railways and high-speed trains. The findings can provide reference for decision-makers to make practical and effective decisions from the perspective of passenger travel preferences.