The invention discloses a traffic flow prediction method and system based on non-embedded spatio-temporal feature fusion, and aims to obtain a high-precision traffic flow prediction result based on a non-embedded spatio-temporal feature integration mode, reduce the dependence of a model on a graph and improve the model training efficiency. The method specifically comprises the steps that different from an existing prediction model based on a space-time diagram convolutional neural network, time features and space features are integrated through a non-embedded method, time sequence information in historical features is fully mined through different time sequence relation capturing methods, dependence of the prediction model on construction of a space node relation diagram is reduced, and the prediction accuracy is improved. Even if the spatial relationship capture module is cancelled, the prediction effect of the prediction model is not obviously reduced, so that the generalization performance and the practical application value of the model in different traffic networks are improved. Meanwhile, by adopting a non-embedded spatial-temporal feature fusion mode, the complexity of the model can be obviously reduced, and the training efficiency can be improved.

    本发明公开了一种基于非嵌入式时空特征融合的交通流预测方法和系统,基于非嵌入式的时空特征整合方式,获得高精度的交通流预测结果,减小模型对图的依赖并提高模型训练效率。具体为:不同于现有的基于时空图卷积神经网络的预测模型,采用非嵌入式的方法来整合时间特征和空间特征,利用不同的时序关系捕捉方法来充分挖掘历史特征中的时序信息,减小预测模型对空间节点关系图的构造的依赖,即使取消空间关系捕捉模块,预测模型的预测效果也不会出现明显下降,从而提高模型在不同交通网络中的泛化性能和实际应用价值。同时,采用非嵌入式的时空特征融合方式,还能明显减小模型的复杂度并提高训练效率。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Traffic flow prediction method and system based on non-embedded spatial-temporal feature fusion


    Weitere Titelangaben:

    一种基于非嵌入式时空特征融合的交通流预测方法和系统


    Beteiligte:
    ZHENG JIANYING (Autor:in) / LUO YONG (Autor:in) / CHEN QIAN (Autor:in)

    Erscheinungsdatum :

    17.09.2024


    Medientyp :

    Patent


    Format :

    Elektronische Ressource


    Sprache :

    Chinesisch


    Klassifikation :

    IPC:    G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G06Q Datenverarbeitungssysteme oder -verfahren, besonders angepasst an verwaltungstechnische, geschäftliche, finanzielle oder betriebswirtschaftliche Zwecke, sowie an geschäftsbezogene Überwachungs- oder Voraussagezwecke , DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS



    Full-period traffic flow prediction method based on spatial-temporal feature deep fusion

    ZHAO ZHONGNAN / XIE XU / WANG YUE | Europäisches Patentamt | 2025

    Freier Zugriff

    Traffic flow prediction method and device based on local-global spatial-temporal feature fusion

    ZONG XINLU / CHEN ZHEN / WANG CHUNZHI et al. | Europäisches Patentamt | 2023

    Freier Zugriff

    Spatial-temporal feature fusion traffic flow prediction method and device, equipment and storage medium

    SHEN QING / ZHANG XIANGZHENG / LOU JUNGANG et al. | Europäisches Patentamt | 2023

    Freier Zugriff

    Spatial-temporal feature enhancement method and system for few-sample traffic flow prediction

    YAO MENGMENG / WANG LULU / CAO YUEHAO et al. | Europäisches Patentamt | 2025

    Freier Zugriff

    Spatial-Temporal Multiscale Fusion Graph Neural Network for Traffic Flow Prediction

    Hou, Hongxin / Ning, Nianwen / Shi, Huaguang et al. | IEEE | 2022