Traffic Accident Analysis on US 158: Machine Learning and HSM Comparison
Published:Dec 26, 2025 03:42
•1 min read
•ArXiv
Analysis
This paper applies advanced statistical and machine learning techniques to analyze traffic accidents on a specific highway segment, aiming to improve safety. It extends previous work by incorporating methods like Kernel Density Estimation, Negative Binomial Regression, and Random Forest classification, and compares results with Highway Safety Manual predictions. The study's value lies in its methodological advancement beyond basic statistical techniques and its potential to provide actionable insights for targeted interventions.
Key Takeaways
- •Applies advanced statistical and machine learning methods to analyze traffic accidents.
- •Identifies spatial and temporal crash patterns on US 158.
- •Random Forest classifier predicts injury severity with 67% accuracy.
- •Validates and extends earlier hotspot identification methods.
- •Provides actionable insights for improving traffic safety.
Reference
“A Random Forest classifier predicts injury severity with 67% accuracy, outperforming HSM SPF.”