Analysis on crash reduction factors for road segment safety

The CRF (crash reduction factor), which has credibility that can quantitatively explain effects on traffic accidents in the field of traffic safety, is an important area which can evaluate the safety of roads. In this research, the influence on traffic safety using accident data was investigated, focusing on road factors and surrounding environmental factors from road engineers’ point of view. In this study, the predictive levels of CRFs, which were used to evaluate the safety of roads by quantitatively expressing effects on traffic accidents, were developed. Accident models were developed by using data from roadways. It was shown that rather than Poisson or negative binomial, zero-inflated Poisson was more appropriate when the frequency in which 0 (zero) case occurs is high. It is also shown that the horizontal curve and radius, whether a concave section exists, the number of lighting systems, whether level terrain or mountainous terrain exists, and the number of crosswalks have influence on accident in certain sections of road.