Heart Disease Prediction Using Machine Learning with Recursive Feature Elimination for Optimized Performance

Abstract

Heart disease remains one of the leading causes of mortality worldwide, emphasizing the need for early diagnosis and effective predictive models to improve patient outcomes. This research focuses on developing a machine learning-based predictive model for heart disease detection by integrating Recursive Feature Elimination (RFE) as a feature selection technique. The primary objective is to enhance model performance, interpretability, and computational efficiency by eliminating irrelevant and redundant features while retaining the most significant predictors. The study evaluates three machine learning algorithms—Logistic Regression, Random Forest, and Support Vector Machine (SVM)—with and without RFE to assess the impact of feature selection on performance. The methodology involves data preprocessing, including normalization and scaling, followed by RFE-based feature selection and model evaluation using key metrics such as accuracy, precision, recall, F1-score, and ROC-AUC. Experimental results reveal that the Random Forest model achieves the highest performance, with 99% accuracy and a 1.00 ROC-AUC, making it the most reliable model for predictive tasks. However, the RFE-based Logistic Regression model provides better interpretability and reduced complexity, albeit with slightly lower performance metrics. The findings highlight the effectiveness of RFE in optimizing feature selection while validating the trade-offs between accuracy and model transparency. This research contributes to the growing field of healthcare analytics by demonstrating the feasibility of using feature selection techniques for building interpretable, scalable, and accurate models for heart disease prediction. Future work will focus on expanding the dataset, incorporating temporal data, and exploring hybrid models to further improve predictive performance and clinical applicability