Enhanced Brain Stroke Prediction: A Bayesian Network-Integrated Ensemble Approach

Stroke remains a predominant global health issue, underscoring the critical necessity of sophisticated predictive models that can elevate diagnostic accuracy and optimize therapeutic interventions. In this study, we propose an advanced stroke prediction framework that leverages an ensemble of Logistic Regression, Random Forest, XGBoost, and Bayesian Network classifiers, meticulously integrated to improve predictive performance. Among the classifiers, the Bayesian Network model achieved a remarkable accuracy of 99.78% on the "brain_stroke" dataset, demonstrating superior performance over the other ensemble models and highlighting the efficacy of probabilistic graphical models in stroke prediction. The proposed methodology offers a robust, high- accuracy diagnostic tool, facilitating early stroke detection, timely clinical intervention, and enhanced resource allocation, thereby contributing significantly to improved patient outcomes and healthcare efficiency.