Efficient Model for Early Prediction of Heart Disease Using Ensemble Technique

The growing global burden of cardiovascular diseases has created an urgent need for advanced early-detection devices that revolutionize preventive cardiology. This research presents a novel two-stage ensemble (TSE) learning framework that outperforms traditional machine learning methods by integrating multiple complex algorithms, including random forest, adaptive boosting, gradient boosting machine, light gradient boosting machine, and extra trees classifier, into a highly accurate predictive system in stage 1. The approach incorporates a sophisticated preprocessing pipeline with feature scaling and the synthetic minority oversampling technique SMOTE to address the class imbalance and ensure robust input data quality. The model optimizes a meta-learner for enhanced predictions by leveraging meta-features derived from various classifiers. The developed TSE model, utilizing the CatBoost classifier in stage 2, achieved average accuracies of 92.5% and 90.19% on the Cleveland and Statlog datasets, respectively. This comprehensive ensemble framework significantly advances clinical decision support for early detection and intervention in cardiovascular disease.