Exploring an Alternative Method of Hazardous Location Identification: Using Accident Count and Accident Reduction Potential Jointly

Abstract

As one of the initial steps to improve the overall safety within a road network, hot spots identification (HSID) has been the focus of many research efforts. Over the past 60 years a wide range of HSID methods have been proposed and applied. One dimension of diversity among these methods lies within the selection of one of two different criteria: the number of accidents or the accident reduction potential. Currently, there is a lack of conclusive evidence showing which criterion performs better. To avoid prematurely favoring one or the other, the authors explore the joint use of both criteria to conduct HSID. New evaluation indexes are proposed and employed. The empirical and full Bayesian estimates and the typical estimate are explored. The evaluation is conducted based on the accident data of two types of Arizona road sections. Additionally, four different sizes of hot spots are assessed. By using the proposed evaluation indexes, the results illustrate that when used in conjunction this new method significantly outperforms the methods using the individual criterion. The findings indicate the advantage of the combination in screening the hot spots with higher accident counts and reduction potentials.

Comparative evaluation for alternative variable importance rankings for pedestrian injury severities

“…Recently, with the rapid advancement in technology and computational capabilities, many variable importance ranking methods have been proposed to facilitate the selection of significant covariates feeding into the subsequent model development. The multitude of techniques contain log-likelihood ratio test [9], partial least squares [10], recursive partitioning [11], bagged trees and boosted trees, Random Forest (RF) [12], Principal Component Analysis (PCA) [13], and some others. The safety profession has witnessed the applications of some of the methods in the past.…”