Impacts of nongeometric attributes on crash prediction at urban signalized intersections of developing countries

Abstract

Urban crashes in the signalized intersections have become a worsening road safety problem and result in unacceptably high socioeconomic toll placing a heavy burden on health and economics. Few studies have been conducted to incorporate the nongeometric attributes in generating crash prediction model for developing countries. Hence, this research investigates the effects of such attributes, that is, traffic volume, bus passenger’s activities, on-street parking, establishments nearby intersection, on the probability of crashes in urban signalized intersections of developing countries through crash prediction models (CPMs) under Empirical Bayesian (EB) framework. Mean absolute error, root mean square error, average accuracy, and statistical correlation of crash count have been utilized for the validation of the proposed models. Total road crashes are likely to be increased by around 5.3% for 10% increase in traffic volume at the intersection whereas presence of on-street parking on major roads is causing a 34.2% reduction in predicted total collision. Low bus passenger activity has been found as a relatively higher effect on crashes than the high bus passenger activity. Such findings have the potential to provide new insights in the occurrences of crashes and quantify the effects of several unique nongeometric attribute at urban intersections.