Predict road accidents resulted from carelessness using negative binomial regression model (A case study of rural highways in Hamadan Province)

Document Type : Original Article

Authors

1 department of engineering, payame noor university, tehran

2 Civil Engineering Department, Malayer University

Abstract

Road accidents are one of the most important causes of mortality in Iran as severe road injuries, their financial, social and economic consequences threaten our society and the humanity as a whole. Rural road accidents compared to urban road accidents are more damaging and due to high speeds often result in very heavy casualties and financial loss. Road accidents are also the major cause of injuries in the world, and the findings of the World Health Organization show that rural road accidents represent 25% of losses worldwide. In this study, using accident data, we try to predict accidents caused by careless driving in certain area of the main corridors of Hamadan province (in IRAN). At first, the factors affecting the accidents were identified, and then we developed a road accident model using a negative binomial regression method based on the maximum probability approach. We made use of the traffic data and geometric design of 945 km of the Hamadan province road network in four axes of Hamadan-Malayer, Hamadan-Asadabad, Hamadan-Saveh and Hamadan-Kabodarahang(in IRAN) from 2012 to 2015. Out of 1145 accidents recorded in the study area, 980 accidents in accident black spots were randomly selected from a total of 534 accidents caused by careless driving and were selected for use in 97 segments in the model. As the final model was developed and verified by goodness of fit tests, we specified the data to be used in the model and outside the model. The results showed that the indicator of proximity to population centers has the most effect on accidents resulted from the carelessness, and factors such as the slope and curve, traffic crash time, length of segments and slopes are the next effective factors.

Keywords

Main Subjects