Prediction Of The Duration Of Highway Traffic Accident

In recent years,with the rapid growth of highway mileage,the frequent accidents of Highway accidents are becoming more and more serious.On the one hand,traffic accidents can cause different levels of traffic jams,increase travel time,increase travel costs,and reduce traffic efficiency of the whole highway network at the same time.More important is the traffic accidents threaten the safety of person and property,easily lead to a series of social problems.Therefore,the rapid rescue of traffic accidents and the timely guidance of traffic congestion are of great significance to the efficient and safe operation of the highway.Among them,predicting the duration of the traffic accident timely and accurately is the precondition of effective implement traffic control,and can provide evidence for releasing of traffic information timely which is induced and predictive as well as eliminating the influence of accident quickly.The duration of traffic accident can be divided into four parts:accident detection time,accident response time,accident clearance time and traffic recovery time.Usually,the first three parts are studied as accident delay time.This paper is based on the analysis of the factors affecting the delay time of highway traffic accidents,respectively on delay time and traffic accident time to build a prediction model,it realizes the estimation of traffic accidents caused by traffic accidents in different traffic conditions and traffic conditions,as well as the overall description of the change of traffic flow in the accident.The research results of this paper mainly include the following aspects:(1)The Decision Tree(DT)model based on C4.5 algorithm is constructed to predict the delay time of highway traffic accidents.Based on the comprehensive data content description and the significance analysis of the influencing factors,11 indicators were established as the attribute layer of decision tree,and 36 time interval categories were divided.On this basis,the decision tree was generated using C4.5 algorithm,and the structure optimization was carried out through the Pessimistic Pruning algorithm.It is proved that the decision tree model based on the attribute selection and classification interval of this paper has a good adaptability to the prediction of accident delay time of different types and forms.(2)The Decision Tree method is a kind of inductive learning algorithm based on examples,considering that the accident happens in a timely manner for comprehensive accident information is less likely,therefore,in order to solve the problems that the accident information is missing or incomplete,construct a Bayes Decision Tree,(Bayes Decision Tree,BDT)to improve the model.On the basis of the optimal DT,the model adds Bayesian nodes.By using this node to determine whether the corresponding attribute information is known,the simple Bayesian theory is used to discriminate the classification of the missing attributes.The results show that the improved model is closer to the actual situation,and the prediction of different data loss is better.(3)This paper is on the theory of traffic flow wave model of highway traffic flow,through the analysis of traffic flow rate and occupancy change in highway accidents,discusses the mechanism of diffusion and dissipation of traffic congestion and illuminates the road traffic capacity in different stages of the impact on traffic recovery time.Though introducing the team to drive vehicles in the concept of virtual possession and building a model of traffic flow wave based on the dynamic space share,it has realized the accident vehicle under the maximum queue length and the shortest queue dissipation time of prediction,and the example shows that the improved model has better prediction effect.