Analysis Of Causes And Severity Of Freeway Traffic Accidents Based On Time And Space

In recent years,the freeway has developed rapidly and the traffic volume has increased rapidly.The potential safety hazards of traffic accidents that follow cannot be ignored.In the related research on the analysis of causes and severity of freeway traffic accidents,the existing research mainly focuses on a certain administrative area or a certain topographical feature from a spatial perspective,such as a general study of freeways in a province,and a single study of freeways in mountainous areas.This lacks a comprehensive analysis of different terrains,ignoring that the characteristics of freeways under different terrains are different,and the severity of accidents caused by the same accident form or cause is not the same.In the perspective of time,weekdays and weekends are usually considered as an influencing factor,and are less regarded as research objects for accident analysis.Therefore,this paper divides the study area into three categories: plains,mountainous areas and hills,and further refines the time into weekdays and weekends.The causes and severity of freeway traffic accidents under different time and space conditions are studied respectively.Taking the freeway traffic accidents in Hebei Province as the research object,the main research contents are as follows:(1)Analysis of the causes of freeway traffic accident under different time and space conditionsAccording to the characteristics of traffic accident data,a multi-dimensional and multilayered attribute structure with 4 dimensions including time,road,environment and accident,and 11 attributes including hour,road alignment,weather,and accident form is built.Adopting the improved Apriori algorithm considering the orientation constraints of accident attributes,the weekdays and weekends of plains,mountainous areas,hills,and a total of 6 different time and space conditions of freeway accidents are analyzed from all cause angles,accident attribute autocorrelation angles,and directional mining angles of specific attribute and accident attribute,focusing on association rules with high support and high confidence.The results show that different association rules are obtained from time angle and space angle,and freeway traffic accidents under different time and space conditions have different influencing factors and coupling mechanisms.Research shows that under different time and space conditions,the accuracy of the improved Apriori algorithm is increased by 86.8%-88.4%,which is suitable for the identification of risk factors of freeway traffic accidents and can reveal the differences in the causes of freeway accidents under different time and space conditions.(2)Analysis of the severity of freeway traffic accidents under different time conditionsA total of 13 variables are selected from the three dimensions of vehicle,road,and environment,including the number of vehicles involved,the type of roadside protection facilities,and the lighting conditions.For accidents on weekdays and weekends in plains,the ordered Logit model and multiple Logit models are applied to construct freeway traffic accident severity models.The models are calibrated by the maximum likelihood estimation method,and the likelihood ratio test,Pearson’s chi-square statistics,deviation statistics and information criterion statistics are carried out to test fitting effects of the models,and the prediction accuracy of two types of models are compared.The results show that freeway traffic accidents under different time conditions have different risk factors,and different models have their own advantages and disadvantages for different time conditions and different severity of accidents.Research shows that the ordered Logit model is more suitable for the analysis of the severity of freeway accidents in plains,and can reveal the differences in the causes and severity of freeway accidents under different time conditions.