Scholarlink

<p>In edge computing environments, accurate and timely Point of Interest (POI) recommendation is critical for enhancing tourist user experience and service efficiency. The distinct characteristics of edge computing, such as the mobility of users and the dynamic nature of spatial data, pose challenges to traditional tourist recommendation systems. This paper addresses these challenges by treating historical interaction data at various locations and time slots as a temporal sequence of POI matrices. We innovate by integrating Autoregressive Integrated Moving Average (ARIMA) and Singular Value Decomposition (SVD). This integration allows for the effective compression of large-scale spatial-temporal data while enabling the ARIMA model to predict POI preferences simultaneously and accurately. Experimental results demonstrate that our proposed approach significantly outperforms existing state-of-the-art methods in both prediction robustness and efficiency while guaranteeing prediction accuracy even in the sparse data environment. This improvement supports more robust and scalable POI recommendations, facilitating better decision making in mobile and dynamic user environments.</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

Optimized tourist point-of-interest recommendation through ARIMA and SVD in edge environment

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月12日

查看全文
  • Haibo Dong,
  • Zhenxiang Liu,
  • Edris Khezri

摘要

In edge computing environments, accurate and timely Point of Interest (POI) recommendation is critical for enhancing tourist user experience and service efficiency. The distinct characteristics of edge computing, such as the mobility of users and the dynamic nature of spatial data, pose challenges to traditional tourist recommendation systems. This paper addresses these challenges by treating historical interaction data at various locations and time slots as a temporal sequence of POI matrices. We innovate by integrating Autoregressive Integrated Moving Average (ARIMA) and Singular Value Decomposition (SVD). This integration allows for the effective compression of large-scale spatial-temporal data while enabling the ARIMA model to predict POI preferences simultaneously and accurately. Experimental results demonstrate that our proposed approach significantly outperforms existing state-of-the-art methods in both prediction robustness and efficiency while guaranteeing prediction accuracy even in the sparse data environment. This improvement supports more robust and scalable POI recommendations, facilitating better decision making in mobile and dynamic user environments.