Research On Identification Of Accident Black-spots On Rural Roads With Clustering Analysis

The construction and development of rural roads have become the key emphasis in work in the decisive stage of building a well-off society in an all-round way.The Ministry of transport and many provincial and municipal governments have proposed the goal of building the "four good rural roads",which indicates the traffic safety of rural roads is increasingly valued.At the same time,the construction and transformation of rural roads in China speeded up rapidly,leading the direct identification method which uses historical data on traffic accidents as the basis of evaluation will not be suitable for the current black-spots identification.Based on the research of accident black spots at home and abroad,a black-spots identification mechanism based on cluster analysis is proposed in this paper.Firstly,in view of the characteristics of rural highways which are different from expressways,a road clustering method based on grid clustering is established.Secondly,principal component analysis is used to reduce the dimension of potential black spots.Based on the improved K-means,cluster analysis of principal component scores is conducted,SSE and Silhouette Coefficient are used as evaluation criteria for clustering results.By using the road accident data,the level of the accident road unit is determined according to the relative size of the comprehensive loss degree.The clustering results show that the critical index of the accident point is scientific.Furthermore,on the basis of the identified black spots,this paper focuses on the statistical characteristics of traffic management data records.In this paper,the improved fuzzy clustering is proposed to explore the main factors,secondary factors,inducement factors,hidden factors and negligible factors,and the Xie-Beni index is used to evaluate the clustering effect.The results show that the improved fuzzy clustering has better accuracy and noise immunity.The data of rural highway accidents in a county in central Jiangsu Province in 2017 were analyzed.The accident black spots were identified,the main reasons were explored,and the improvement for main reasons were analyzed according to the actual situation.Comparing the physical truth,accuracy of the accident black spot analysis mechanism are verified.The paper provides a theoretical basis for the regulation of accident black spots in traffic safety management of rural roads.