The invention discloses a traffic accident risk research and judgment method based on an interpretable random forest. The method comprises the following steps: acquiring traffic accident data and road section data of a research area, and preprocessing the data; according to the method, an urban road traffic accident risk influence index system is constructed from the perspective of three elements including traffic flow factors, road factors and environmental factors, an improved random forest-based urban road traffic accident risk research and judgment model is constructed, the model is explained by adopting an SHAP method, and the safety value range of influence factors is calculated. The method solves the problems that road traffic accident black spots are difficult to identify, accident risks are difficult to assess, accident causes are difficult to analyze, and an accident influence mechanism is unclear to explain, research, judgment and cause analysis of urban road traffic accident risks are achieved, calculation errors are small, convergence speed is high, efficiency is high, interpretability is high, and the method is suitable for popularization and application. And a theoretical basis is laid and technical support is provided for road traffic accident risk research and judgment and dynamic analysis of risk causes.

    本发明公开了一种基于可解释随机森林的交通事故风险研判方法,包括:获得研究区域的交通事故数据和路段数据并对数据进行预处理;从交通流因素、道路因素及环境因素三要素的角度构建城市道路交通事故风险影响指标体系,构建基于改进随机森林的城市道路交通事故风险研判模型,采用SHAP方法对模进行解释,计算影响因素的安全取值范围。本发明解决了道路交通事故黑点难以鉴别,事故风险难以评估,事故致因分析困难,事故影响机理解释不清的难题,实现了城市道路交通事故风险的研判与致因分析,计算误差较小且收敛速度快,效率较高,有较强的可解释性,为道路交通事故风险研判以及风险致因的动态分析奠定理论基础并提供技术支持。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Traffic accident risk studying and judging method based on interpretable random forest


    Weitere Titelangaben:

    一种基于可解释随机森林的交通事故风险研判方法


    Beteiligte:
    WAN YUJIE (Autor:in) / DONG CHUNJIAO (Autor:in) / LI PENGHUI (Autor:in) / XU BO (Autor:in)

    Erscheinungsdatum :

    2024-05-28


    Medientyp :

    Patent


    Format :

    Elektronische Ressource


    Sprache :

    Chinesisch


    Klassifikation :

    IPC:    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 / G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS



    Apparatus and method for judging traffic accident

    LEE JAE HYUN / SHIM HWA JIN / KIM YUN HO et al. | Europäisches Patentamt | 2017

    Freier Zugriff

    Traffic accident detection using random forest classifier

    Dogru, Nejdet / Subasi, Abdulhamit | IEEE | 2018


    Expressway special section risk situation studying and judging method based on cumulative effect

    YE QING / LUO ZHONGBIN / WANG SHAOFEI et al. | Europäisches Patentamt | 2023

    Freier Zugriff

    System for judging vehicle accident and performing risk control based on edge calculation

    LI ZHI / WANG SHENG / WEN XIN | Europäisches Patentamt | 2023

    Freier Zugriff

    DEVICE AND METHOD FOR PREVENTING TRAFFIC ACCIDENT THROUGH JUDGING ROAD CONDITION BASED ON DEEP LEARNING MODEL

    JINSUL KIM / DONGSU LEE / ASHIQUZZAMAN AKM et al. | Europäisches Patentamt | 2021

    Freier Zugriff