Time series forecasting of road accident in Malaysia by using Box-Jenkins and univariate model / Nur Farah Bahiah Mohd Zarmi, Nur Balqish Zainal and Norliana Mohd Lip

This study aims to predict future road accidents in Malaysia using time series forecasting techniques. The objectives are to determine the best model between Box-Jenkins and Univariate methods, and to forecast road accident numbers based on the model with the lowest error measures of MAPE and RMSE. The study analyzes Malaysia’s historical and current traffic accident trends to provide insights for forecasters, government agencies, and transportation firms. Researchers can also build upon the findings to further investigate traffic accident prediction in Malaysia. The results show that the Double Exponential Smoothing model provides the most accurate forecasts for 2024 road accident numbers in Malaysia. This information can help authorities implement targeted interventions to improve road safety and reduce the significant human and economic toll of traffic accidents.