基于多项Logit模型的机动车相关事故影响因素和严重程度探究  

Research on Influencing Factors and Severity of Motor Vehicle-related Crashes Based on Multinomial Logit Model

作　　者：王天文 庄越[1,2] WANG Tianwen;ZHUANG Yue(School of Safety Science and Emergency Management,Wuhan University of Technology,Wuhan 430070,China;China Research Center for Emergency Management,Wuhan University of Technology,Wuhan 430070,China)

机构地区：[1]武汉理工大学安全科学与应急管理学院,武汉430070 [2]武汉理工大学中国应急管理研究中心,武汉430070

出　　处：《交通工程》2023年第2期11-18,共8页Journal of Transportation Engineering

基　　金：国家社会科学基金项(20BGL252).

摘　　要：为研究涉及机动车事故严重程度的综合影响因素,收集了美国2020年机动车事故数据,从驾驶员、环境、道路、事故特征4个方面选取28个自变量,事故严重程度为因变量,采用多项Logit、有序Logit、部分比例优势模型进行对比,选择最佳模型(多项Logit)对各因素进行量化分析.结果表明驾驶员年龄、性别,是否酒驾、超速、无证驾驶,不良天气和照明条件、道路线形、路面状态、事故时段、交叉路口类型、是否使用安全带、事故地区、事故类型、事故位置、碰撞方式和车辆类型对事故严重程度有显著性影响.最后基于显著影响因素有针对性地为交管部门提出改进措施.In order to study the comprehensive influencing factors related to the severity of motor vehicle crashes,this paper collected the data of motor vehicle crashes in the United States in 2020,took the severity of crashes as dependent variables,selected 28 independent variables from four aspects:driver,environment,road conditions and crash characteristics,compared them with multinomial Logit,ordered Logit and proportional odds models,and identified the best model(multinomial Logit)to quantitatively analyze each factor.The results showed that the driver s age,gender,drunk driving,speeding,driving without a license,bad weather,lighting conditions,road alignment,road condition,crash time,intersection type,seatbelt use,crash area,crash type,crash location,collision type and vehicle type had significant influence on the severity of the crash.Finally,based on the significant influencing factors,some improvement measures are put forward for the traffic management department.

关 键 词：机动车事故 LOGIT模型 定量分析 交通安全对策 

分 类 号：U491[交通运输工程—交通运输规划与管理]