T Manouchehri - Department of Statistics, Shiraz University, Shiraz, Iran
S.T Heydari - Health Policy Research Center, Shiraz University of Medical, Shiraz, Iran
K Bagheri Lankarani - Health Policy Research Center, Shiraz University of Medical, Shiraz, Iran
R Fereidooni - Health Policy Research Center, Shiraz University of Medical, Shiraz, Iran

Traffic accidents globally pose a significant public health crisis, resulting in loss of life, injuries, and substantial economic burdens. Predictive analytics, combining data science and health policy, offers a solution by identifying accident factors and creating proactive safety measures to save lives. This study uses machine learning to analyze traffic accident data from self-report questionnaires, revealing the importance of socio-economic status, psychosocial factors, and long driving hours in predicting accident occurrence. The research suggests interventions targeting socio-economic factors, psychological assessments and improved road safety to effectively prevent accidents.

Traffic accidents; Predictive analytics; Machine learning; Feature selection