Scholarlink

<p>The objective of this research is to better understand the heterogeneity of passengers' travel behavior and their spatial locations, which can be used to optimize public transport network structure. Based on individual passenger’s space-behavior analysis from a two-dimensional perspective, this paper first used a spatial nearest neighbor search algorithm to divide passengers into a number of passenger groups within a predefined space. Then, the diversified public transport markets were identified by subdividing these groups into several sub-groups with similar activities through clustering index screening and the Mini-Batch K-means clustering algorithm. In addition, the clustering results were also compared with other clustering algorithms such as DBSCAN, and it was found that the Mini-batch K-means algorithm was relatively fast and effective. Finally, the rationality of this method was verified by using an actual case in Xiamen City, China, and some reasonable and practical operation suggestions were further provided for optimizing and adjusting the specific transit network.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Public transport market segmentation based on space-behavior analysis

  • 期刊论文

  • 发表日期：

    2025年12月12日

查看全文
  • Linbo Li,
  • Huajian Gao,
  • Jianan Fu,
  • Yahua Zhang,
  • Ziyuan Li

摘要

The objective of this research is to better understand the heterogeneity of passengers' travel behavior and their spatial locations, which can be used to optimize public transport network structure. Based on individual passenger’s space-behavior analysis from a two-dimensional perspective, this paper first used a spatial nearest neighbor search algorithm to divide passengers into a number of passenger groups within a predefined space. Then, the diversified public transport markets were identified by subdividing these groups into several sub-groups with similar activities through clustering index screening and the Mini-Batch K-means clustering algorithm. In addition, the clustering results were also compared with other clustering algorithms such as DBSCAN, and it was found that the Mini-batch K-means algorithm was relatively fast and effective. Finally, the rationality of this method was verified by using an actual case in Xiamen City, China, and some reasonable and practical operation suggestions were further provided for optimizing and adjusting the specific transit network.