The age information of urban rail transit passengers is of great significance for accurate passenger flow organization and personalized intelligent service. Aiming at the problems of missing and difficult to obtain the passenger age attribute of rail transit, this paper integrates multi-source data mining such as automatic fare collection (AFC) data and urban land use data to reflect the travel characteristics of passenger age, and proposes a passenger age attribute inference model combining deep neural network (DNN) and automatic encoder (AE). Guangzhou Metro is selected for case analysis. The results show that compared with the baseline model SVM, MLP, DT and AdaBoost, the accuracy of DNN+AE model proposed in this paper is improved by 11.08%, 9.82%, 5.33% and 3.33%, respectively. Among them, the accuracy of age attribute inference for the elderly is the highest, reaching 78.58%.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    An Age Attribute Inference Method for Urban Rail Transit Passengers Based on Deep Learning


    Contributors:
    Liu, Jing (author) / Xu, Xinyue (author) / Zhang, Anzhong (author) / Ye, Hongxia (author)


    Publication date :

    2023-09-24


    Size :

    608374 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Spatiotemporal path inference model for urban rail transit passengers based on travel time data

    Luo, Qin / Lin, Bin / Lyu, Yitong et al. | Wiley | 2023

    Free access


    Role of Passengers in a Modern Urban Rail Transit System

    Xu, Xiaofan / Shen, Liping / Li, Hong | ASCE | 2013


    Study on Route Choice Behavior of Urban Rail Transit Passengers

    Liu, Sha Sha ;Yao, En Jian ;Zhang, Yong Sheng | Trans Tech Publications | 2014


    Travel trajectory detection of travelers in urban rail transit based on reference passengers

    Zhao, Yafeng / Zhao, Peng | British Library Conference Proceedings | 2022