期刊文献+

考虑时间稳定性的公交事故严重性影响因素分析 被引量:2

Analysis of Factors Influencing Bus Crash Severity Considering Temporal Stability
下载PDF
导出
摘要 为了探究公交事故严重性影响因素的时间稳定性和潜在未观察到的异质性,构建了随机阈值随机参数分层有序probit模型(RTRPHOPIT),并利用2016—2019年的公交事故数据,分析了驾驶员特征、道路环境特征和事故特征变量对死亡、重伤以及轻伤事故的潜在影响。基于似然比检验对影响因素时间稳定性的判断,构建了两个不同时间段的RTRPHOPIT模型,分析了模型参数和阈值的随机性,并对变量的边际效应进行了对比。结果表明:通过允许模型阈值与估计参数具有随观测值变化的随机性,RTRPHOPIT模型能有效地捕捉未观测的异质性,且能充分揭示影响公交事故严重性的因素特征。此外,影响公交事故严重性的因素存在明显的时间不稳定性,不同时间段模型中阈值都是随机分布的,且模型参数和阈值随机性的解释变量存在差异性。其中,侧面碰撞事故和城区道路变量会增加模型的阈值,环形交叉口、多车事故、停车过程事故以及上午事故变量为模型的随机参数。模型变量的边际效应分析结果表明,2016—2017年模型中,乡村驾驶员、夜间无路灯道路、湿滑路面、晴天事故、侧面碰撞事故以及正面碰撞事故变量会增加严重伤害事故发生的可能性;2018—2019年模型中,男性驾驶员、主干道、公交专用道、夜间亮灯道路、翻车事故以及正面碰撞事故变量会增加严重伤害事故发生的可能性。除正面碰撞事故变量外,其他因素变量对事故严重性的影响随着时间的推移而变化。因此,管理者在制定公交安全政策时,需要慎重考虑事故严重性影响因素的异质性及其在时间维度的不稳定性。 We constructed a random thresholds random parameters hierarchical ordered probit model(RTRPHOPIT)to analyze the temporal stability and potential unobserved heterogeneity of factors influencing the severity of bus crashes.Using bus crash data between 2016 and 2019,we analyzed the potential impact of driver characteristics,road environment characteristics,and crash characteristic variables on fatal,serious,and slight injuries.We constructed two RTRPHOPIT over different periods to analyze the randomness of model coefficients and thresholds and compared the marginal effects of the model variables.We based this analysis on the likelihood ratio test’s assessment of the temporal stability of the influencing factors.The results show that by allowing thresholds and parameters to have randomness that vary in different observations,the RTRPHOPIT model can effectively capture the unobserved heterogeneity and reveal the characteristics of factors affecting bus crash severity.Besides,the factors that affect bus crash severity have apparent temporal instability.The model thresholds for different periods are randomly distributed.The random parameters and the explanatory variables that affect threshold values are inconsistent.Specifically,side-impact crashes and urban road variables will increase the threshold value of the model.The variables,including roundabouts,multi-vehicle crashes,parking process crashes,and morning crashes,are the random parameters of the model.The marginal effect of the 2016-2017 model shows that some variables(e.g.,rural drivers,night-without-light roads,wet slippery roads,fine-day crashes,side-impact crashes,and front-impact crashes)will increase the likelihood of serious injury crashes.Similarly,other variables in the 2018-2019 model(e.g.,male drivers,major roads,bus lanes,night-light roads,rollover accidents,and front-impact crashes)will increase this possibility.Therefore,when formulating bus safety policies,authorities must carefully consider the temporal instability of the factors affecting the severity of the bus crash,and pay considerable attention to the heterogeneity of model variables at the same time.
作者 沈金星 刘坤 于淼 马昌喜 SHEN Jin-xing;LIU Kun;YU Miao;MA Chang-xi(College of Civil Engineering and Transportation Engineering,Hohai University,Nanjing 210098,China;School of Traffic and Transportation,Lanzhou Jiaotong University,Lanzhou 730070,China)
出处 《交通运输工程与信息学报》 2022年第1期108-118,共11页 Journal of Transportation Engineering and Information
基金 国家自然科学基金项目(51808187) 中央高校基本科研业务费专项资金资助项目(B210202035)。
关键词 城市交通 事故严重性 随机阈值随机参数分层有序probit模型 公交事故 时间稳定性 urban traffic crash severity random thresholds random parameters hierarchical ordered probit model(RTRPHOPIT) bus crash temporal stability
  • 相关文献

参考文献4

二级参考文献30

共引文献49

同被引文献13

引证文献2

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部