Logo image
Back
A MULTI-FACTOR COMBINED MODEL FOR PREDICTING TRAFFIC CONFLICT SEVERITY
Conference paper

A MULTI-FACTOR COMBINED MODEL FOR PREDICTING TRAFFIC CONFLICT SEVERITY

Hailin Kang, Hongkang Song, Shaohu Tang and Liang Zhang
Proceedings of the 2025 International Annual Conference and 46th Annual Meeting: Powering the Future of Engineering Management, ASEM 2025, pp 572-581
2025

Abstract

Job market analysis for Systems Engineers Skills required for Systems Engineers (SE) Systems Engineering (SE)
With the increasing number of motor vehicles in cities, intersections—key nodes for traffic flow—have become more complex and diverse. Therefore, predicting the severity of traffic conflicts at intersections is crucial for preventing traffic accidents. This paper investigates both signalized and unsignalized cross-shaped urban intersections, utilizing a combination of video and manual surveys to extract traffic conflict parameters. The factors influencing traffic conflict severity are analyzed using an ordered multi-class logistic regression model. Based on these factors, a combined model for predicting traffic conflicts severity is developed by integrating Genetic Algorithm (GA), Support Vector Machine (SVM), Backpropagation (BP) Neural Network, and Random Forest (RF). The model’s performance is evaluated using accuracy, precision, recall, and F1-score. The results show that the GA-RF model outperforms the other two combination models in prediction accuracy and effectively predicts traffic conflict severity at urban intersections. This predictive capability serves as a technical foundation for dynamic risk assessment and proactive safety measures.

Metrics

1 Record Views

Details

Logo image