Machine Learning Approach to Predict Dengue Fever in Bangladesh

Abstract

Although Bangladesh is endemic to Dengue fever and has been reporting outbreaks of the disease, it still contin ues to remain an important public health concern. This study developed and evaluated multiple machine learning models to predict dengue fever using an open-access clinical hematology dataset comprising 2,086 samples sourced from the Mendeley Data repository. The models evaluated include Random Forest, Logistic Regression, Decision Tree, Support Vector Machines (SVM), K-Nearest Neighbors (KNN), GaussianNB, XGBoost, and Stacking Ensemble. The study results indicated that Ran dom Forest achieved the highest accuracy (88.73%), followed closely by the Stacking Ensemble (88.01%) and XGBoost (87.05%). Random Forest and Logistic Regression performed the best in terms of minimizing false positives and false negatives, suggesting that the use of such models may be beneficial for more efficient surveillance monitoring and timely interventions in controlling Dengue Fever in Bangladesh.