.png){kind=icon}
.png){kind=icon}
Return
Exploring transit bus related fatal crash patterns at intersections and roadway segments using association rule mining
.png){kind=icon}
PRE
AI
S
S
A
T
S
DOI：10.1080/12265934.2026.2684750.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Transit buses are vital for safe, sustainable urban mobility, yet fatal crashes involving these vehicles remain a critical concern for city planners and policymakers. This study applies Association Rule Mining (ARM) with the Lift Increase Criterion (LIC) to examine multiyear patterns of fatal transit bus crashes in the United States from 2016 to 2023, using data from the Fatality Analysis Reporting System (FARS). The analysis distinguishes between intersection-level and segment-level environments, revealing context-specific risk profiles shaped by driver demographics, roadway design, lighting conditions, and traffic controls. At intersections, weekday crashes often involved elderly and middle-aged drivers under dark-lighted conditions on county roads, while summer crashes on U.S. highways were frequently linked to younger drivers and rear-end impacts. Segment-level crashes were associated with winter conditions, uncontrolled or rural roadways, and frontal collisions among drivers aged 45–64. Additional rules highlighted risks for young male drivers in dense urban areas and weekday pedestrian crashes during daylight hours. By applying LIC, the study identifies high-strength co-occurrence patterns that traditional models often obscure, offering actionable insights for targeted interventions. The findings underscore the need for improved lighting, stronger traffic control, roadway redesign, and driver-focused safety programs. These measures directly support Vision Zero and other urban safety frameworks, providing evidence-based pathways to reduce fatalities and strengthen the reliability of transit within cities.
Keywords:
Transit bus crashes
association rule mining
roadway segments
intersections
lift increase criterion
Journal
I
IF:
3
Papers:
463
Citations:
1.1K
Organization
Cited Papers