Modeling and forecasting trafic accident severity based on main factors

Abstract

The purposes of this study were to construct main factors causing traffic accidents by measuring attributes of an accident, and to model the ordered regression relationship between those factors and the severity of the accident. The result showed that the hazard factor due to speed and collision direction and the risk factor due to control difficulty were two main factors that significantly affect the severity of a traffic accident. In validating the model, we found that within 1 unit diffirence, the model showed an accuracy of 82.8, which means that the model is viable for practical application to estimate the severity of accidents and to help delivering preventive measures.