《计算机应用》唯一官方网站 ›› 2024, Vol. 44 ›› Issue (7): 2065-2072.DOI: 10.11772/j.issn.1001-9081.2023071045

• 数据科学与技术 • 上一篇    下一篇

基于多尺度时空图卷积网络的交通出行需求预测

李欢欢, 黄添强(), 丁雪梅, 罗海峰, 黄丽清   

  1. 福建师范大学 计算机与网络空间安全学院,福州 350108
  • 收稿日期:2023-08-02 修回日期:2023-09-25 接受日期:2023-10-09 发布日期:2023-10-26 出版日期:2024-07-10
  • 通讯作者: 黄添强
  • 作者简介:李欢欢(1997—),女,贵州遵义人,硕士研究生,主要研究方向:深度学习、时空数据挖掘;
    黄添强(1971—),男,福建莆田人,教授,博士,主要研究方向:人工智能、多媒体内容安全、大数据分析;
    丁雪梅(1972—),女,黑龙江哈尔滨人,教授,博士,主要研究方向:机器学习、异常检测、数据科学、智能计算;
    罗海峰(1990—),男,福建龙岩人,讲师,博士,主要研究方向:三维点云智能处理、时空数据挖掘;
    黄丽清(1991—),女,福建莆田人,讲师,博士,主要研究方向:视频图像处理、多媒体内容安全、人工智能。
  • 基金资助:
    国家自然科学基金资助项目(62072106);福建省自然科学基金资助项目(2022J01188);福建省教育厅中青年教师教育科研项目(JAT210051)

Public traffic demand prediction based on multi-scale spatial-temporal graph convolutional network

Huanhuan LI, Tianqiang HUANG(), Xuemei DING, Haifeng LUO, Liqing HUANG   

  1. College of Computer and Cyber Security,Fujian Normal University,Fuzhou Fujian 350108,China
  • Received:2023-08-02 Revised:2023-09-25 Accepted:2023-10-09 Online:2023-10-26 Published:2024-07-10
  • Contact: Tianqiang HUANG
  • About author:LI Huanhuan, born in 1997, M. S. candidate. Her research interests include deep learning, spatial-temporal data mining.
    HUANG Tianqiang, born in 1971, Ph. D., professor. His research interests include artificial intelligence, multimedia content security, big data analytics.
    DING Xuemei, born in 1972, Ph. D., professor. Her research interests include machine learning, anomaly detection, data science, intelligent computing.
    LUO Haifeng, born in 1990, Ph. D., lecturer. His research interests include intelligent processing of 3D point clouds, spatial-temporal data mining.
    HUANG Liqing, born in 1991, Ph. D., lecturer. Her research interests include video image processing, multimedia content security, artificial intelligence.
  • Supported by:
    National Natural Science Foundation(62072106);Fujian Natural Science Foundation(2022J01188);Educational Research Project for Young and Middle-aged Teachers of Fujian Provincial Department of Education(JAT210051)

摘要:

满足公众高质量出行需求是智能交通系统(ITS)的主要挑战之一。目前,针对公共交通出行需求预测问题,现有模型大多采用固定结构的图描述出行需求的空间相关性,忽略了出行需求在不同尺度下具有不同的空间依赖关系。针对上述问题,提出一种多尺度时空图卷积网络(MSTGCN)模型。该模型首先从全局尺度和局部尺度构建全局需求相似图和局部需求相似图,这2种图可以捕获公共交通出行需求长期内较为稳定的全局特征和短期内动态变化的局部特征。利用图卷积网络(GCN)提取2种图中的全局空间信息和局部空间信息,并引入注意力机制融合两种空间信息。为了拟合时间序列中潜藏的时间依赖关系,利用门控循环单元(GRU)捕捉公共交通需求的时变特征。采用纽约市出租车订单数据集和自行车订单数据集进行实验,结果表明MSTGCN模型在自行车订单数据集上均方根误差(RMSE)、平均绝对误差(MAE)和皮尔逊相关系数(PCC)达2.788 6、1.737 1、0.799 2,在出租车订单数据集上RMSE、MAE、PCC达9.573 4、5.861 2、0.963 1。可见,MSTGCN模型可以有效地挖掘公共交通出行需求的多尺度时空特性,对未来公共交通出行需求进行准确预测。

关键词: 公共交通出行需求预测, 图卷积网络, 时空数据挖掘, 注意力机制, 深度学习, 智能交通系统

Abstract:

High-quality public traffic demand has become one of the major challenges for Intelligent Transportation Systems (ITS). For public traffic demand prediction, most of existing models adopt graphs with fixed structure to describe the spatial correlation of traffic demand, ignoring that traffic demand has different spatial dependence at different scales. Thus, a Multi-scale Spatial-Temporal Graph Convolutional Network (MSTGCN) model was proposed for public traffic demand prediction. Firstly, global demand similarity graph and local demand similarity graph were constructed at global and local scales. Two graphs were used to capture long-term stable and short-term dynamic features of public traffic demand. Graph Convolutional Network (GCN) was introduced to extract global and local spatial information in two graphs; besides, attention mechanism was adopted to combine the two kinds of spatial information adaptively. Moreover, Gated Recurrent Unit (GRU) was used to capture time-varying features of public traffic demand. The experimental results show that the MSTGCN model achieves the Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and Pearson Correlation Coefficient (PCC) of 2.788 6, 1.737 1, and 0.799 2 on New York City (NYC) Bike dataset; and 9.573 4, 5.861 2, and 0.963 1 on NYC Taxi dataset. It proves that MSTGCN model can effectively mine multi-scale spatial-temporal features to accurately predict future public traffic demand.

Key words: public traffic demand prediction, Graph Convolutional Network (GCN), spatial-temporal data mining, attention mechanism, deep learning, Intelligent Transportation System (ITS)

中图分类号: