Cardiovascular Disease Prediction Using Ensemble Stacking for Enhanced Accuracy

Cardiovascular disease remains a noteworthy global health concern, warranting accurate prediction methods for early diagnosis and intervention This research investigates the application of fundamental machine learning models for predicting Cardiovascular Disease (CVD) hazards while exploring the effectiveness of ensemble stacking techniques to enhance predictive accuracy. Basic models including Random Forest, Decision Tree, XG Boost,Gaussian Naïve Bayes, and Light GBM were trained and evaluated on a comprehensive dataset containing patient attributes and clinical indicators related to CVD. To solve the issue of class imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) was utilized. Subsequently, an ensemble stacking approach was employed to combine predictions from these models, demonstrating its ability to boost predictive performance. Results showcase the superiority of the ensemble stacking method, yielding a remarkable accuracy of 98.4%. This study highlights the comparative performance of basic machine learning models and underscores the potency of ensemble techniques in significantly improving predictive accuracy for CVD risk assessment, emphasizing their potential in proactive healthcare management.