Description of Invention
Traffic accidents remain a major road safety concern, requiring accurate severity prediction for effective emergency response. Traditional models often miss complex factor interactions, reducing real-world impact. This study applies machine learning to classify accident severity—minor, serious, or fatal—using Random Forest, Gradient Boosting, and Logistic Regression based on features like road conditions, weather, and driver profiles. Random Forest performed best, achieving 58.9% accuracy, 64.7% precision, and 35.9% recall. Key factors included driver age, experience, and environment. The proposed system offers a data-driven approach to aid traffic authorities in emergency preparedness and resource allocation, improving response efficiency and road safety strategies.