BIMSA > BIMSA Digital Economy Lab Seminar Impact of real-time weather conditions on crash injury severity in Kentucky using the correlated random parameters logit model with heterogeneity in means
Impact of real-time weather conditions on crash injury severity in Kentucky using the correlated random parameters logit model with heterogeneity in means
Organizers
Li Yan Han , Zhen Li , Fei Long , Ke Tang , Yu Wang
Speaker
Meiyi Wang
Time
Monday, January 20, 2025 3:20 PM - 4:20 PM
Venue
A3-2-303
Online
Zoom 230 432 7880 (BIMSA)
Abstract
This study examines the influence of real-time weather conditions—such as air temperature, relative humidity, precipitation, wind speed, and solar radiation—on the severity of crash injuries. The research merges crash data from January 2016 to April 2021 on Interstate-75 in Kentucky with weather information from the Kentucky Mesonet stations at a one-hour granularity.To assess the effect of various weather conditions on crash severity, the study introduces a novel severity index (SI), which compares the ratio of severe crashes to the exposure of specific weather conditions during the crash period. The study also applies several advanced statistical models, including the standard mixed logit (MXL), correlated mixed logit (CMXL), and correlated mixed logit with heterogeneity in means (CMXLHM), to account for unobserved heterogeneity and identify the risk factors contributing to crash injury severity.The findings reveal that the CMXLHM model outperforms the other models in terms of statistical fit, as indicated by metrics such as the Akaike information criterion (AIC) and McFadden’s pseudo R-squared. Key results from both the SI analysis and the CMXLHM model show that weather factors such as air temperature (≥ 70°F) and high relative humidity (≥ 90%) are significantly associated with higher likelihoods of severe injuries in crashes. Additionally, factors like driving under the influence (DUI), young drivers, and vehicle speed are linked to greater injury severity, while factors such as the presence of horizontal curves, passenger cars, and traffic volume contribute to lower injury severity likelihood.
Speaker Intro
Meiyi Wang is a PhD student at BIMSA and UCAS.Her research interests focus on digital economy, international finance, and AI-driven driving technologies.