Journal of Transportation Systems Engineering and Information Technology ›› 2017, Vol. 17 ›› Issue (6): 126-132.

Previous Articles     Next Articles

High-speed Railway Passenger Ticketing Behavior Characteristics Based on PCA and Clustering

LIU Fan-xiao a, b, PENG Qi-yuan a, b, LIANG Hong-bin a, FU Zhi-jian a, ZHANG Bin a   

  1. a. School of Transportation and Logistics; b. National United Engineering Laboratory of Integrated and Intelligent Transportation, Southwest Jiaotong University, Chengdu 610031, China
  • Received:2017-06-20 Revised:2017-09-11 Online:2017-12-25 Published:2017-12-25

基于PCA-聚类分析的高铁旅客购票行为特性研究

刘帆洨a, b,彭其渊*a, b,梁宏斌a,傅志坚a,张斌a   

  1. 西南交通大学a. 交通运输与物流学院;b. 综合交通运输智能化国家地方联合工程实验室,成都610031
  • 作者简介:刘帆洨(1985-),女,重庆长寿人,工程师,博士生.
  • 基金资助:

    国家自然科学基金/National Natural Science Foundation of China (U1234206, 61571375);中央高校基本科研业务专项计划资助项目/Fundamental Research Funds for the Central Universities(2682013BR021).

Abstract:

The ticketing behavior is an important basis for the railway transportation operation planning. It influences the occupation process of the train capacity. Based on the ticketing statistics data from Gui-Guang high- speed railway, principal component analysis (PCA) is used to analyze the behavior characteristics comprehensively to obtain the critical characteristic variables, while the ticketing passengers are considered as the samples. Considering the tickets probability density function for a single purchase, the ticketing passengers are clustered by double clustering based on fuzzy C-means algorithm and the optimal number of cluster K is determined by fuzzy clustering effectiveness index: Xie- beni and Separation Coefficient. The result shows that the number of passengers per trip, the number of ticketing days in advance, GDP per capital of original-destination (OD) cities and pre-sell approach are the significant factors. The ticketing behaviors preformed distinctively in different passenger types.

Key words: railway transportation, ticketing behavior, PCA, fuzzy C-means, double clustering, ticketing passenger

摘要:

铁路运输中旅客购票行为是铁路客运运营策略制定的重要基础.旅客购票行为直接影响着列车能力的占用过程,是铁路客运票额组织的重要依据.根据贵阳—广州高速铁路的旅客购票统计数据,以高铁购票旅客为样本,运用主成分分析法购票行为的特征属性进行综合分析,获得购票行为中重要特征变量.结合单次购票强度提出基于模糊C均值的双重聚类算法,对购票旅客进行聚类,并利用模糊聚类有效性指标Xie-beni 和分离系数法确定最佳聚类数.结果表明,高铁旅客购票行为的关键特性为单次出行旅客人数、购票提前天数、出行OD城市人均GDP和购票渠道;不同旅客类型的购票行为有明显特性.

关键词: 铁路运输, 购票行为, 主成分分析, 模糊C 均值, 双重聚类, 购票旅客

CLC Number: