Abstract
The objective of this study is to clarify a relationship between the road traffic conditions and accident frequency through statistical model analysis towards implementing effective safety measures. A rich accident database was combined with relevant road traffic information by using GIS software, and two different models, i.e. negative binomial regression and multiple regression models were independently applied for estimating accident frequency. As a result, it was found that accident frequency is closely related to average travel speeds and distance between intersections in arterial roads. Regarding non-arterial roads, the result specified different influencing factors by accident type. Furthermore, it suggested a high potential of frequent accident occurrence in non-arterial roads particularly at positions close to the arterial roads with low average travel speed. Finally, the necessity of improving quality of traffic flow and hierarchical road network for road safety was discussed.
