Prediction of vehicle traffic accidents using bayesian networks

Every year, thousands of vehicle accidents occur in Iran and result thousands of deaths, injuries and material damage in country. Various factors such as driver characteristics, road characteristics, vehicle characteristics and atmospheric conditions affect the injuries severity of these accidents. In order to reduce the number and severity of these accidents, their analysis and prediction is essential. Currently, the accidents related data are collected which can be used to predict and prevent them. New technologies have enabled humans to collect the large volume of data in continuous and regular ways. One of these methods is to use Bayesian networks. Using the literature review, in this study a new method for analysis and prediction of vehicle traffic accidents is presented. These networks can be used for classification of traffic accidents, hazardous locations of roads and factors affecting accidents severity. Using of the results of the analysis of these networks will help to reduce the number of accidents and their severity. In addition, we can use the results of this analysis for developing of safety regulations.