Predicting Road Accidents in Bani Walid, Libya, Using Neural Network Artificial Intelligence

Abstract

Road safety is one of the most essential elements that must be present on any road, as it is concerned with preserving the lives of road users, as well as protecting private and public property, reducing financial costs, and saving time and effort. This research aims to improve traffic safety by attempting to predict the future pattern of accidents using artificial intelligence and to provide an idea and an early warning if the current accident rate continues in the study city (Bani Walid). The available accident data from the city’s traffic center for the past twelve years (2013–2024) was analyzed, organized, and then used to feed machine learning models in MATLAB after training the models in order to forecast the number of accidents in the future for the years 2025–2030. The focus was placed on the most important accident statistics indicators: ''number of fatalities, serious injuries, and number of damaged vehicles''. The results indicate a noticeable overall increase in future statistics, especially in fatalities, where the model, based on previous patterns, predicted that the number of fatalities will continue to rise, exceeding 50 deaths per year, with 65 serious injuries and 110 damaged vehicles. This highlights the urgent need to implement road safety strategies immediately.