عنوان مقاله [English]
Speed limit signs are posted alongside roads, based on their geometric characteristics, except for driver behavior and possible environmental conditions, to reduce traffic hazards. This study summarizes the results obtained by 527 questionnaires that were given to the drivers of vehicles moving on the Chaboksar-Ramsar road, north of Iran, from December to January. About 9.5\% of the drivers are female. The research is of a theoretical-practical type. The data is collected by random selection in statistics. The drivers answered questions voluntarily for about ten minutes. First, we did descriptive statistics. In the second step, using a classification and regression tree (CART) technique, the most important factors affecting driver behavior, and attracting their attention to the 50km/hr speed limit sign, among the 16 such factors identified, including age, gender, education, income, marital status, occupation, driving speed, driving time, driving hours, vehicle type, number of accidents over the previous 5 years, driving certificate type, driving experience, the years of driving certificate taken, and sunglasses used, were found to be driver education, monthly income, and type of vehicle. The CART is a nonparametric model and capably used for categorical variables in various sciences. The parameters are known from most past studies. The results indicated that three parameters are the most important (not that other parameters are unimportant) in driving behavior. To validate the results, a speed limit sign near the study location was used. The speeds of a sample of 414 vehicles were observed by the speedometer installed on the asphalt pavement at this location. The results show that the drivers do not pay sufficient attention to the posted speed limit sign. We suggest that researchers evaluate other models in the same zones and compare the results.