Research On The Prediction And Evaluation Method Of Urban Traffic Safety

With continuous progress of urban modernization has been made, urban road infrastructure and vehicle population have been improved substantially, and people's living standards have been ameliorated a lot. However these achievements have been made, a series of road traffic safety problems were came into being at the same time. How to predict and evaluate urban traffic safety has become a serious problem to be solved urgently.On the basis of purposes and principles of urban road traffic safety prediction and evaluation, the factors which affect urban traffic safety were discussed from four aspects, such as human being, vehicle, road and environment, and a urban road traffic safety prediction index system was created, whose results'relative indicator contained accident rate per 10000 cars, accident rate per 10000 capita, mortality per 10000 cars, mortality per 10000 capita and so forth. As well as the creation of urban traffic safety prediction index system, a evaluation index system was crated,which contained road traffic safety facilities, traffic flow status, the traffic accident situation, traffic management and safeguard and the traffic environment aspects.After the characteristics of traffic safety prediction and evaluation have been discussed respectively, and the advantages of using Support Vector Machine theory and Extenics in urban traffic safety prediction and evaluation have been described, the models of urban traffic safety prediction and evaluation based on Support Vector Machine theory and Extenics were established.Finally, as an example of Hefei, prediction and evaluation models based on SVM and Extenics were established separately using Hefei traffic safety data which was collected from 2006 to 2009, and models'feasibility and effectiveness were verified by comparing and analyzing.