Afaneh, Nura (University of Calgary), Demissie, Merkebe Getachew (University of Calgary, Calgary, Canada), Kattan, Lina (University of Calgary)

Traffic Collision Occurrence and Severity Prediction for Enhanced Clearance Time

Scheduled for presentation during the Regular Session "Traffic prediction and estimation I" (WeAT6), Wednesday, September 25, 2024, 11:30−11:50, Salon 14

2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), September 24- 27, 2024, Edmonton, Canada

This information is tentative and subject to change. Compiled on October 14, 2024

Abstract

Road transportation remains the primary contributor to traffic collisions, resulting in property damage, injuries, and fatalities. Previous efforts were mainly directed at understanding the causes of traffic collisions and predicting their occurrence. This study also examines the issue of traffic collisions, focusing on predicting the occurrence and severity of traffic collisions. We also extended our effort to predict collision clearance time, essential for optimizing emergency response and advanced road transportation management. We applied several traditional statistical and machine learning techniques, including multinomial logistic regression, ridge regression, K-nearest neighbor, support vector machine, random forest, and neural network. All aspects of the study were conducted using collision data from Calgary, Alberta, Canada, spanning from 2022 to 2024. The study finds that while traditional regression models offer good initial predictions, machine learning techniques like random forest and neural networks demonstrate superior predictive accuracy. By addressing the spatiotemporal nature of traffic collision prediction and using advanced analytical techniques, this research contributes valuable insights into resource allocation for enhancing road safety.