Speeding has been distinguished as one of the most frequent and persistent contributing factors and is a critical contributing factor to the degree of injury severity. In the United States, at least a quarter of nationwide annual fatal crashes during the last decade involved speeding. There is still a need for an overarching look at crashes involving speeding by considering a wider set of crashes, roadway, driver, and vehicle characteristics. Despite extensive research on speeding-related crashes, there is limited understanding of how collective variables in homogeneous crash clusters contribute to fatal speeding crashes. This paper addresses this gap by investigating these collective impacts using fatal crash data from the Fatality Analysis Reporting System (FARS). Using crash data from the 2015–2019 FARS repository, this study applies latent class clustering (LCC) to obtain homogeneous clusters of fatal speeding crashes, addressing the unobserved heterogeneity. Association rule mining (ARM) has been applied to homogeneous clusters to find hidden patterns. The finding of association rules, such as motorcycle speeding, single vehicle crashes during weekends, in dark, unlit conditions, etc. The results of this research and interpretative findings are expected to improve the knowledge of speeding-related crash mechanisms and to provide important insights on countermeasure development.
Identifying attribute associations in fatal speeding crashes using latent class clustering and association rule mining
M. A. Rahman et al.
Journal of Transportation Safety & Security
Journal of Transportation Safety & Security ; 17 , 5 ; 510-549
2025-05-04
40 pages
Article (Journal)
Electronic Resource
English
Causal Insights into Speeding Crashes
ASCE | 2024
|Fractions of fatal crashes attributable to speeding: Evolution for the period 2001–2010 in France
Online Contents | 2013
|