Crash Severity

Crash severity research aims to understand and predict the outcome of traffic accidents, focusing on factors influencing injury levels and fatalities. Current research heavily utilizes machine learning, employing algorithms like random forests, gradient boosting, and deep neural networks, often coupled with techniques to address imbalanced datasets (e.g., synthetic data generation) and improve model interpretability (e.g., SHAP values). These advancements offer improved accuracy in predicting crash severity and identifying key contributing factors, informing the development of more effective road safety strategies and potentially improving autonomous vehicle safety systems.