Predicting traffic accidents on Tehran's highways using discrete selection models

Traffic accidents are a social problem in the country that kills a large number of people every year and imposes huge economic costs on society. In recent years, much attention has been paid to traffic management through accident prediction methods. Predicting an accident has a great effect on reducing casualties, injuries and financial losses caused by accidents. The purpose of this study is to determine the share of each of the effective factors in the occurrence of accidents by introducing a model to predict the occurrence of accidents on highways in Tehran, different areas of Tehran. This research is a applied type and in terms of data collection method, it is a descriptive survey. The statistical population of the present study was the statistics and data related to accidents on the highways of Tehran in 2018, which have been registered in the I.R.I. Police's Traffic Police Database. Due to the fact that all the reports related to the accidents in Tehran have been investigated. Therefore, limited sampling is avoided and all available data are analyzed. In this study, discrete selection model (logit) has been used to predict the causes of accidents. The findings of this study show that the most important causes of accidents on Tehran's highways have been due to inattention to the front, non-observance of longitudinal distance, sudden change of direction, non-observance of the right of way. Vehicles leaving the parking lot, speeding, and riding on the road have increased the number of such accidents. The results showed that traffic accidents on the highways of Tehran can be predicted by the Logit model in the coming years and the factors affecting them can be determined and controlled.