Feature elimination and stacking framework for accurate heart disease detection in IoT healthcare systems using clinical data.

Abstract

Heart disease remains a complex and critical health issue, necessitating accurate and timely detection methods. In this research, we present an advanced machine learning system designed for efficient and precise diagnosis of cardiac disease. Our approach integrates the power of Random Forest and Ada Boost classifiers, along with incorporating data pre-processing techniques such as standard scaling and Recursive Feature Elimination (RFE) for feature selection. By leveraging the ensemble learning technique of stacking, we enhance the model's predictive performance by combining the strengths of multiple classifiers. The evaluation metrics results demonstrate the superior accuracy and obtained the higher performance in terms of accuracy, 99.25%. The effectiveness of our proposed system compared to baseline models. Furthermore, the utilization of this system within IoT-enabled healthcare systems shows promising potential for improving heart disease diagnosis and ultimately enhancing patient outcomes.