Prioritization of Effective Factors in Rural Road Accidents of Guilan Province Based on Exploratory Factor Analysis and Logistic Regression Model

Nowadays, it is necessary to understand the factors affecting accidents, especially on rural roads. In this study, in order to determine the most effective factor in the occurrence of accidents, analysis and modeling of effective factors on rural roads of Guilan province in 1392 to 1397 have been conducted. First, frequency analysis was used to evaluate the variables and their frequency. Then, Friedman test was utilized to prioritize the factors and exploratory factor analysis was used to determine the most effective factor in the occurrence of vehicle accidents. Finally, multiple logistic regression models were used to predict the probability of the occurrence of accidents. Based on Friedman test results, weather conditions, road surface conditions and the type of vehicle accident were identified as the first to third factors affecting accidents, respectively. Exploratory factor analysis showed that five factors as the main factors are involved in accidents that the variables of weather conditions and road surface conditions as an environmental factor were recognized as the first effective factor in accidents. Also, multiple logistic regression models with more accuracy in predicting the severity of accidents (84.7%) showed that cloudy weather, dry road surface and wet road surface had the highest effect on occurrence of accidents, respectively. Results of accident sensitivity analysis also showed that logistic regression (with the area under the curve 0.932) was much more accurate than factor analysis (with the area under the curve 0.849), which shows the high power of this model in predicting and evaluating of accidents severity.