高峰期城市道路交通事故严重程度影响因素分析

Analysis of Factors Influencing the Severity of Urban Road Traffic Accidents during Peak Period

利用国内某市2016—2020年的历史交通事故数据,探究高峰期城市道路交通事故严重程度的影响因素。首先,定量描述事故数据的分布和特征,选取驾驶人年龄等14个因素作为候选自变量,以事故严重程度作为因变量,利用逐步向前法和似然比检验筛选显著影响因素。研究发现,车辆类型、车辆安全状况、道路线形、路侧防护设施、天气5个因素与高峰期事故严重程度显著相关。然后,构建二项Logistic回归模型,进行模型系数的Omnibus检验和模型Hosmer-Lemeshow检验。结果表明,模型拟合程度良好;摩托车、轻型货车、大中型客车、重型货车更容易发生严重事故;车辆安全状况不佳时发生严重交通事故概率大;非平直线形道路发生严重交通事故的概率大于平直线形道路;路侧无防护不易发生严重交通事故;阴雨天相较于晴天发生的交通事故更严重。基于模型分析结果,提出了道路交通管理的相关建议,为减轻高峰期交通事故严重程度提供了依据和参考。

The factors influencing the severity of urban road traffic accidents during peak hours was investigated by using the historical traffic accident data from a city in China from 2016 to 2020.Firstly,the distribution and characteristics of the accident data were quantitatively described.14 factors such as driver's age were selected as candidate independent variables,with accident severity as the dependent variable.Significant influencing factors were screened out by stepwise forward method and likelihood ratio test.It was found that vehicle type,vehicle safety condition,road alignment,roadside protective facilities,and weather were significantly correlated with accident severity during traffic peak period.Subsequently,a binary logistic regression model was established,and the Omnibus test and the Hosmer-Lemeshow test of the coefficients of the model were performed.The results exhibited a good fit for the model;severe accidents were more likely to involve motorcycles,light trucks,medium and large-sized buses,and heavy trucks;the probability of severe accidents on non-straight roads was higher compared to straight roads;roads without protection were less likely to present severe accidents;and accidents that occurred during cloudy and rainy weather were more severe than in sunny weather.Based on the model analysis results,relevant recommendations for road traffic management were proposed,which provided a basis and reference for alleviating the severity of traffic peak period accidents.