Enhancing Dengue Outbreak Prediction in Bangladesh: A Comparative Study of Advanced Predictive Models

Dengue fever, a major global health challenge caused by the dengue virus (DENV) and transmitted by Aedes mosquitoes, affects nearly half of the world’s population, with 100 to 400 million annual infections. This study focuses on predictive modeling to forecast dengue hospital admissions in Bangladesh, aiming to enhance early intervention and resource allocation. Using daily reported cases from the Directorate General of Health Services (2019-2023), three models were evaluated: Seasonal Autoregressive Integrated Moving Average (SARIMA), Extreme Gradient Boosting (XGBoost), and Random Forest Regressor (RFR). SARIMA outperformed the others with the lowest Mean Squared Error (MSE: 0.3011) and highest R2 (0.9418), demonstrating its effectiveness in capturing seasonal trends. This model shows significant potential for improving dengue surveillance and outbreak management in Bangladesh. The findings underscore the importance of accurate predictive modeling in mitigating dengue outbreaks and optimizing public health responses.