Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
May 31, 2026Quality & QuantityOpen Access

Interpretable and robust tree-based methodology for imbalanced classification in driving safety assessment

View Full Paper
Ask AI
Bookmark
Share

Authors

GVGiulia VannucciADA D’AmbrosioRSRoberta Siciliano

Discussion

Loading...

Member takes

Overview

Randomized trial identifies causal relationships influencing safety-critical driving outcomes, suggesting improved risk assessment methods.

Key Points

  • The study aims to uncover causal relationships between human behaviors, external factors, and safety-critical driving events.
  • Analyzed the SHRP2 naturalistic driving database for context-specific driving situations.
  • Employed a three-step framework incorporating multiple correspondence analysis and classification trees.
  • Utilized random forests and global sensitivity analysis to evaluate variable importance in driving safety.
  • Identified latent behavioral typologies influencing the likelihood of crashes or near-crashes.
  • Demonstrated that contextual factors significantly affect the risk of safety-critical driving events.
  • Provided actionable insights for improving road safety through a transparent and interpretable framework.

Cite This Study

Vannucci et al. (2026) studied this question.

synapsesocial.com/papers/6a1bd1f65783ba022b6fd58fhttps://doi.org/10.1007/s11135-026-02877-w
View Full Paper
Ask AI
Bookmark
Share