Poisson Log-Linear Regression Model for Rural Signalized Intersection

摘要: Safety performance of rural signalized intersections is critical for identifying high-risk sites and predicting the hazardousness. This paper aims to develop a predictive model that will describe safety based on various input variables. Data are examined from 124 over three states, Poisson log-linear regression presented, which connected traffic number average volumes, geometric characteristics signalization variables together. The associated data analysis reveal daily traffic, media width, speed limit, degree horizontal curvature left-turn lane factors have greatest overall effect safety. results show able intersection accurately. Using this model, effective countermeasures can be applied improving