AI Assisted Innovation for Predicting Traffic Accident on Addis -Adama Expressway

Abstract

Despite the critical importance of expressway safety, there is a notable absence of a dedicated and accurate predictive model specifically designed to forecast traffic accidents on the Addis-Adama Expressway. My research examines the application of artificial intelligence (AI) to solve the lack of a tailored predictive model and emergency services to proactively address safety concerns on Addis Adama Expressway by analyzing historical traffic accident datasets using various machine learning models, including Random Forest, Support Vector Machine, K-Nearest Neighbor, gradient boosting, and Decision Tree classifiers. My analysis identifies the major factors contributing to road Traffic accidents on the Addis Adama Expressway, such as Chainage, driver age, driver experience, day of the week, cause of accident, geo-location, vehicle types, crash type, weather condition, driver relationship, road surface condition, and sex. And also, I identify support vector machines as the most effective model, achieving an accuracy of 78.11%. Therefore, I developed a road traffic accident prediction model using support vector machines. Furthermore, I recommended that Continuous Awareness Training on Safe Driving to enhance road safety, implement a centralized traffic data system that continuously collects and analyzes information from sources such as GPS, traffic cameras, and sensors to identify accident patterns and high-risk areas, and also use predictive analytics to forecast potential hotspots and deploy preventive measures proactively.