An Advanced Ensemble Framework for Robust Heart Disease Detection and Classification

Abstract

Cardiovascular Disease (CVD) remains one of the leading causes of mortality worldwide, emphasizing the need for accurate and early diagnostic solutions. Recent advances in Machine Learning (ML) and Deep Learning (DL) have shown significant potential to support clinical decision-making through data-driven prediction models. This study presents a robust Ensemble Learning (EL) framework for the prediction and classification of CVD by integrating multiple ML algorithms with a DL component. Specifically, an Artificial Neural Network (ANN) is employed as a feature extraction layer prior to ensemble aggregation using techniques such as Random Forest, XGBoost, and LightGBM. The proposed approach is evaluated using accuracy, precision, recall, F1-score, and AUC-ROC. Experimental results on a benchmark dataset demonstrate that the model achieves a high accuracy of 98.8%, outperforming individual classifiers and existing approaches. The integration of ANN-based feature extraction enhances model generalization and reduces prediction error. These findings highlight the effectiveness of the proposed framework for early heart disease detection and clinical decision support.