Prediction Model And Analysis, Bayesian Network Theory-based Traffic Incident

Road traffic safety is an important issue which restricts the development of modern road traffic system. It not only brings a huge threat to people's lives and property, but also to impose an important influence on society as a whole. The valid prediction of road traffic incidents is the very important contents of road traffic safety research for grasping the development tendency of road traffic incidents in the future,providing scientific basis for decision-making for road traffic management and carrying on the reasonable evaluation on the feasibility of road traffic safety measures and implementation effect that used to predict.On the basic of analyzing these research methods, the traffic event forecast model was researched and analyzed in this dispersion by the Bayesian network method which was used in the field of probability safety assessment widely, and provided new mentality for improving traffic safety conditions of the traffic management department. The paper expounds the relationship between the various factors and the traffic incidents by summarizing and analyzing the various factors which affect traffic safety. And, the dissertation established traffic incidents prediction model by Bayesian network method. In this model, traffic flow,speed,climate and other incidents four factors are considered, and the specific modeling steps and probabilistic reasoning algorithm are given. Based on historical data of traffic condition, it is verified the simulation model, the experimental result is analyzed and the validity of the model is proved.