| The curve or downgrade section is the section with traffic accident prone in the road traffic environment,and it occupies the important position in road traffic. There is so little special study on the traffic accident of the curve or downgrade section by scholars at home and abroad.On the one hand,the study on the traffic accident of the dangerous section that is all focus on the driver factors,and the road factors is in the secondary position;On the other hand,in the study of the influence on traffic accidents based on the road conditions,most is the analyzed of relationship between the microscopic parameters of road design and traffic accidents.In actual road conditions,the microscopic parameters of road design and the driver factors is difficult to obtain.In this paper,on the basis of the fuzzy information of the road accidents,use decision tree method for curve or downgrade section to analyze and predict traffic accident,to provide a valid traffic information for the driver who driving a vehicle traveling on the curve or downgrade section.The main contents of this paper is divided into three parts.The first section has summarized the domestic and international trends accident prediction method and prediction of traffic accidents,and clear the contents and research ideas of this paper.The second part made a detailed analysis of the types of accidents which occured at high frequency on the curved or downgrade section by using the FTA method,and the effect of the road conditions of all types of accidents on the curved section,including the impact of the number of lanes and other road curve radius design conditions for all kinds of traffic accidents,as well as visibility,road conditions and other environmental conditions on all types of road traffic accidents.The third part analysis more than 3,000 from the curved section of the traffic accident data samples by the CART algorithm,to obtain the classification rules of predicted traffic accident that applies curved section,K cross-validation showed a high accident prediction accuracy,and the model can be used.To predict the accident of a case of a curved section of highway,and made some traffic safety recommendations on the basis of forecast results.The research method is easy to understand and implement,able to handle both data type and conventional type attribute,it can be made feasible and the effect of good results for large data sources in a relatively short period of time,and provide technical support for the promoting of the active safety car. |
