Integration of heuristic and statistical methods for estimation of cyclist injury severity

Abstract

The number of crashes that involve bicycles and result in severe injuries has increased every year in the Republic of Korea, though bicycling is a promising alternative transportation mode that can help address transportation emission problems. Correspondingly, this study intends to quantitatively examine the impacts of contributing factors on the cyclist injury severity levels and to provide meaningful insight into prioritizing and specifying strategies for improving cyclist safety. To this end, a decision tree and ordered probit regression were integrated in this study. The findings showed that heavy vehicle use and cyclist age should be preferentially considered when implementing cyclist safety improvement strategies: targeting young heavy-vehicle drivers or cyclists, driver education regarding proper curve maneuvering and a law enforcement prohibition of speeding are suggested, and ITS-based traffic management system and road facilities would help enhance light vehicle driver visibility and senior cyclist alertness, particularly in horizontally curving road sections. These data-driven results could quantitatively support the policy makers or practitioners making decisions about prioritizing cyclist safety improvement strategy implementations and further specifying such strategies.