The invention discloses a traffic flow prediction method based on an adjacent DBSCAN fused time-varying multi-graph convolutional network, and the method comprises the following steps: collecting traffic flow data, and carrying out the preprocessing of the original data; constructing an adjacent DBSCAN clustering algorithm to perform density clustering on the sensor nodes, and dividing the sensor nodes into a plurality of node clusters; generating a time-varying feature matrix, an adjacent matrix, an incidence matrix and a travel intention matrix for the sensor of each node cluster, and dividing a training set and a test set; constructing a time-varying multi-graph convolutional network, wherein the network is composed of a multi-graph spatial feature extraction module and a time-varying feature extraction model; and training the time-varying multi-graph convolutional network by using the training set, and testing the prediction precision of the model by using the test set. According to the method, the adjacent DBSCAN clustering algorithm is constructed to pre-screen the traffic sensor nodes to effectively obtain the node cluster with high relevance, and the dynamic multi-period spatial-temporal characteristics of the traffic flow are effectively extracted through the constructed time-varying multi-graph convolutional network, so that the traffic flow prediction precision between the high-relevance sensor nodes is improved.

    本发明公开一种基于邻接DBSCAN融合时变多图卷积网络的交通流预测方法,包括以下步骤:采集车流量数据,对原始数据进行预处理;构建邻接DBSCAN聚类算法对传感器节点进行密度聚类,划分为多个节点簇;对每个节点簇的传感器生成时变特征矩阵、邻接矩阵、关联矩阵和出行意愿矩阵,并划分训练集和测试集;构建时变多图卷积网络,该网络由多图空间特征提取模块和时变特征提取模型组成;利用训练集训练时变多图卷积网络,并用测试集测试模型的预测精度。本发明通过构建邻接DBSCAN聚类算法对交通传感器节点进行预筛选有效获取关联性高的节点簇,并通过构建的时变多图卷积网络有效提取交通流的动态多周期时空特征,提高了高关联性传感器节点之间的交通流预测精度。


    Access

    Download


    Export, share and cite



    Title :

    Traffic flow prediction method based on adjacency DBSCAN fused time-varying multi-graph convolutional network


    Additional title:

    基于邻接DBSCAN融合时变多图卷积网络的交通流预测方法


    Contributors:
    SHI QUAN (author) / BAO YINXIN (author) / CAO YANG (author) / SHEN QINQIN (author) / YU XIAN (author) / ZHANG TENGYUN (author) / CAO CHENYANG (author)

    Publication date :

    2023-03-21


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / 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



    Traffic flow prediction method based on time-varying fusion graph convolutional network

    ZUO KAIZHONG / WANG RUI / WANG CHEN et al. | European Patent Office | 2024

    Free access


    Traffic flow prediction method based on graph convolutional network

    XU HUI / MENG FANYU / REN QIANQIAN et al. | European Patent Office | 2025

    Free access

    Space-time ARMA graph convolutional network traffic flow prediction method

    CAO YANG / XIAO PEICHENG / SHEN QINQIN et al. | European Patent Office | 2024

    Free access

    Temporal Multi-Graph Convolutional Network for Traffic Flow Prediction

    Lv, Mingqi / Hong, Zhaoxiong / Chen, Ling et al. | IEEE | 2021