基于二元Logit模型的直角事故受伤严重程度分析

Analysis of the Severity of Right-angle Accident Injury Based on a Dual Logit Model

直角事故是一种常见的交通事故类型,为探寻直角事故受伤严重程度的显著影响因素,本文运用二元Logit模型,从人、车、路和环境四个方面对美国密歇根州Wayne县的直角事故受伤严重程度进行全面分析,且采用向前似然比检验和向后似然比检验两种方法对比得到模型最优拟合优度。研究结果表明:驾驶员因素(未采取紧急措施、不遵守信号灯指示、饮酒)、道路限速、事故中包含摩托车和路面情况是直角事故的显著影响因素,并得到了计算直角事故发生人员伤亡概率的回归方程。本文得到的结论可为交通管理提供参考,以减少该类事故的受伤严重程度。

Right-angle accidents are a common type of traffic accident.To determine the significant influencing factors of the seriousness of injury in such accidents,this study uses the binary logit model to simulate a right-angle accident in Wayne County,Michigan,USA based on four aspects:human,vehicle,road,and environment.The severity of the injury was comprehensively analyzed and the optimal fitness of the model was obtained by comparing the forward and backward likelihood ratio tests.The research results indicate that driver-related factors(not taking emergency measures,not complying with the signal indication,drinking alcohol),road speed limit,accidents including motorcycles,and road conditions are significant factors.The obtained regression equation was used to calculate the probability of non-property damage only(non-PDO)of right-angle accidents.The conclusions obtained herein can be used as a reference for traffic management to reduce the severity of injuries in such accidents.