Scholarlink

<p>This study proposes a hybrid framework to estimate boarding stops in public transit systems using smart card and GPS data, addressing the absence of key fields like vehicle IDs and stop identifiers in entry-only systems. A time-window matching algorithm links transactions to potential stops, followed by passenger classification into high- and low-frequency groups. For frequent travelers, maximum likelihood estimation (MLE) leverages travel chains for accurate inference, while an enhanced k-nearest neighbors (KNN) algorithm predicts stops for infrequent travelers using spatiotemporal proximity. Evaluated on Lanzhou’s Route 80 and Route 77 (June 2023), the method demonstrates high consistency (with cosine similarity exceeding 0.91) and achieves well-balanced passenger flows, with balance coefficients of 0.97 and 0.89. Field surveys confirm alignment with observed patterns at key hubs. This scalable approach is well-suited to urban transit systems with limited data integration, supporting improved planning and operational optimization.</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

Estimating bus boarding stops under missing key fields in smart card transaction data

  • 期刊论文

  • 发表日期：

    2025年10月14日

查看全文
  • Bin Lv,
  • Xianlin Li,
  • Binbin Hao,
  • Qixiang Chen

摘要

This study proposes a hybrid framework to estimate boarding stops in public transit systems using smart card and GPS data, addressing the absence of key fields like vehicle IDs and stop identifiers in entry-only systems. A time-window matching algorithm links transactions to potential stops, followed by passenger classification into high- and low-frequency groups. For frequent travelers, maximum likelihood estimation (MLE) leverages travel chains for accurate inference, while an enhanced k-nearest neighbors (KNN) algorithm predicts stops for infrequent travelers using spatiotemporal proximity. Evaluated on Lanzhou’s Route 80 and Route 77 (June 2023), the method demonstrates high consistency (with cosine similarity exceeding 0.91) and achieves well-balanced passenger flows, with balance coefficients of 0.97 and 0.89. Field surveys confirm alignment with observed patterns at key hubs. This scalable approach is well-suited to urban transit systems with limited data integration, supporting improved planning and operational optimization.