Investigating older driver crashes on high-speed roadway segments: a hybrid approach with extreme gradient boosting and random parameter model

Abstract

Older drivers are often more susceptible to crashes due to age-related physical and cognitive limitations, particularly in complex driving environments. Considering the limited research in this area, this study focuses on investigating crashes involving older drivers on high-speed roadways (≥ 45 mph). The analysis is based on data collected from Louisiana State, encompassing 18,300 older driver-involved crashes (2017-2021). For analysis, a two-step hybrid modelling approach is employed: a) Extreme Gradient Boosting (XGBoost) is used to classify top variable features and b) Correlated Random Parameter Ordered Probit with Heterogeneity in Means (CRPOP-HM) is used to predict the likelihood of crash injury severity. . Some of the critical factors increasing the likelihood of fatal-severe or injury crashes involving older drivers on high-speed segments include the manner of collision (rear-end, right-angle, single-vehicle), primary contributing factor (violation, pedestrian action), presence of passenger (s), location type (open country, residential, business with mixed residential), and weekend.