Scholarlink

<p>With the rapid development of sensor technology and global navigation satellite system (GNSS), the collection of massive and precise spatiotemporal geographic flow data has become increasingly feasible. This study proposes a novel Neighbor Grid-based Adaptive Spatiotemporal DBSCAN (NG-based ASTDBSCAN) algorithm to address key challenges in Origin–Destination (OD) flow clustering, including adaptive parameter optimization and computational efficiency. By integrating grid partitioning, breadth-first search, and adaptive strategies using KNN, KDE, and the ELBOW method, the algorithm significantly enhances clustering speed and accuracy. Experiments on synthetic datasets and real-world Beijing taxi OD flow data demonstrate that the proposed method improves runtime efficiency by up to 28% compared to state-of-the-art algorithms while maintaining high clustering accuracy. The application analysis reveals critical urban travel patterns, offering valuable insights for urban planning and resource allocation. This study provides an efficient and robust solution for largescale spatiotemporal data clustering, with potential for broader applications in geographic flow analysis, urban mobility forecasting and logistics 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

Research on neighbor grid-based adaptive spatiotemporal DBSCAN algorithm

  • 期刊论文

  • 发表日期：

    2026年2月16日

查看全文
  • Haiqi Wang,
  • Yawen Ou,
  • Feng Liu,
  • Xueying Li,
  • Yuanhao Cao,
  • Baozhong Wang,
  • Jun He,
  • Tong Liu,
  • Yanwei Wang

摘要

With the rapid development of sensor technology and global navigation satellite system (GNSS), the collection of massive and precise spatiotemporal geographic flow data has become increasingly feasible. This study proposes a novel Neighbor Grid-based Adaptive Spatiotemporal DBSCAN (NG-based ASTDBSCAN) algorithm to address key challenges in Origin–Destination (OD) flow clustering, including adaptive parameter optimization and computational efficiency. By integrating grid partitioning, breadth-first search, and adaptive strategies using KNN, KDE, and the ELBOW method, the algorithm significantly enhances clustering speed and accuracy. Experiments on synthetic datasets and real-world Beijing taxi OD flow data demonstrate that the proposed method improves runtime efficiency by up to 28% compared to state-of-the-art algorithms while maintaining high clustering accuracy. The application analysis reveals critical urban travel patterns, offering valuable insights for urban planning and resource allocation. This study provides an efficient and robust solution for largescale spatiotemporal data clustering, with potential for broader applications in geographic flow analysis, urban mobility forecasting and logistics optimization.