Modelling The Number of Road Accident Casualities in Southwestern States in Nigeria

Abstract

Road traffic accidents pose a serious public health challenge worldwide, particularly in developing countries where road infrastructure, traffic law enforcement, and safety awareness are often inadequate. This study focuses on traffic fatalities in Nigeria’s southwest region—comprising Lagos, Ogun, Oyo, Ondo, Osun, and Ekiti states—which accounts for a substantial proportion of the country’s vehicular movement and accident burden. Using quarterly traffic accident data from 2013 to 2023 obtained from the Federal Road Safety Corps website, the study applies descriptive analysis, unit root testing, the ARIMA model, and the neural network autoregressive (NNETAR) model to evaluate predictive performance. Model comparison is based on Mean Square Error (MSE), Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE). The results indicate that the NNETAR model outperforms the other approaches by producing the lowest prediction errors across the training datasets, demonstrating its effectiveness in forecasting road traffic casualties in southwest Nigeria. These findings support the use of autoregressive neural network models for traffic accident forecasting and provide valuable insights for improving road safety strategies, while future research may further explore the impact of policy and intervention measures.