Ensemble framework for cardiovascular disease prediction.

Area of Science:

• Cardiology
• Biomedical Informatics
• Machine Learning in Healthcare

Background:

• Heart disease is a leading cause of global mortality, necessitating early and accurate diagnosis.
• Existing diagnostic methods require improvement for timely intervention and improved patient survival rates.
• Machine learning offers promising avenues for developing predictive systems for cardiovascular diseases.

Purpose of the Study:

• To develop and evaluate a machine learning framework for predicting cardiovascular disease risk.
• To leverage a comprehensive dataset combining multiple sources for robust model training.
• To enhance the accuracy and efficacy of heart disease prediction systems.

Main Methods:

• Utilized a large, combined dataset from IEEE Data Port (Hungarian, Cleveland, VA, Switzerland, Statlog).
• Implemented a stacked ensemble classifier integrating ExtraTrees Classifier, Random Forest, and XGBoost algorithms.
• Assessed model performance using metrics including accuracy, ROC, AUC curve, specificity, F1-score, sensitivity, and MCC.

Main Results:

• The proposed stacked ensemble model achieved a high accuracy of 92.34%.
• Performance metrics demonstrated the model's efficacy in predicting cardiovascular disease.
• The developed framework surpassed the accuracy reported in existing literature.

Conclusions:

• The developed machine learning framework shows significant potential for accurate and early heart disease prediction.
• Stacked ensemble methods offer a powerful approach for improving cardiovascular disease risk assessment.
• This research contributes to advancing AI-driven diagnostic tools in cardiology.