Construction and Implementation of a Crash Prediction Model on Two-Way Suburban Two-Lane Roads (Case Study: Qazvin-Avaj Road)

Improving the level of road safety requires research that measures the effects of various variables on the frequency of accidents and takes practical measures to improve the level of road safety. In this study, first, the two inseparable axes of the Qazvin-Avaj suburban line were selected as a case study, and accident information was collected in the mentioned axis for the years 2010 to 2012. The number of accidents was selected as a dependent variable, and the variables of number of access roads and intersections, number of horizontal curves, and hourly traffic volume were selected as independent variables for modeling. The length of the day and night, depending on the spring and autumn hours, was divided into two parts, night and day, and a corresponding model was developed for each of them. In the end, the results were reported in such a way that the importance of effective variables in the day model is: number of access roads and intersections, hourly traffic volume, and then the number of horizontal arches. While for the night model, the variable number of horizontal arches had the most effect, followed by the number of access roads and intersections, and the least effective variable for the night model was the volume of hourly traffic.