Scholarlink

With the acceleration of urbanization, traffic security has become a top priority in the construction of smart cities. Efficient and accurate traffic flow prediction can bring immense convenience to people’s lives, by connecting different networks and optimizing transport systems as a whole. The unique non-linearity and complex spatial-ST-correlation of transport flow data suggest considerable challenges in prediction, as existing spatial-temporal prediction algorithms are based on graph convolution to capture global or heterogeneous relationships, and simpler graph convolution models cannot accurately capture complex spatial relationships. To address the aforementioned issues, this article presents a multi-graph-based spatial-temporal GCN model, built up of different connected graphs to capture fine-grained spatial relationships. Experimental evaluations show that the framework proposed outperforms existing methods with better results in the analysis performed with publicly available datasets and proved that the proposed security prediction model can provide a decision-making basis for solving traffic security problems.

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

A Novel Multi-graph Spatial-Temporal Graph Neural Network Security Model for Traffic Prediction

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Chunyan Diao,
  • Yanrong Zhang,
  • Jinquan Ma,
  • Xi Ouyang,
  • Lixiang Guo,
  • Nan Zhou,
  • Sheng Huo,
  • Tong Shen,
  • Kuanching Li

摘要

With the acceleration of urbanization, traffic security has become a top priority in the construction of smart cities. Efficient and accurate traffic flow prediction can bring immense convenience to people’s lives, by connecting different networks and optimizing transport systems as a whole. The unique non-linearity and complex spatial-ST-correlation of transport flow data suggest considerable challenges in prediction, as existing spatial-temporal prediction algorithms are based on graph convolution to capture global or heterogeneous relationships, and simpler graph convolution models cannot accurately capture complex spatial relationships. To address the aforementioned issues, this article presents a multi-graph-based spatial-temporal GCN model, built up of different connected graphs to capture fine-grained spatial relationships. Experimental evaluations show that the framework proposed outperforms existing methods with better results in the analysis performed with publicly available datasets and proved that the proposed security prediction model can provide a decision-making basis for solving traffic security problems.