Performance Evaluation of Hybrid Machine Learning Models for Heart Disease Prediction in U.S. Clinical Decision Support Systems

Heart-disease continues to be a major contributor to illness and death across the United States. This is highlighting the importance of reliable early-prediction tools. That can aid clinicians in making informed decisions. In this work, researchers analyze the effectiveness of multiple hybrid machine learning approaches for predicting heart disease using a set of clinically meaningful attributes. The evaluated models include Decision Tree (DT), Random Forest (RF), Gradient Boosting Machine (GBM), Support Vector Machine (SVM), and k-Nearest Neighbor (KNN). To enhance model relevance, univariate feature selection was used to determine the most influential predictors. Model performance was measured using accuracy (ACC), precision (PRE), and F1-score. This experimental analysis indicates that the Random Forest algorithm consistently delivered the strongest performance. This is reaching an accuracy of 98.31% on the test set. In addition, limiting the model to the eight highest-ranking features increased interpretability while maintaining high predictive accuracy. Overall, the results suggest that Random Forest both alone and in combination with feature selection provides a robust and dependable framework for heart disease prediction in U.S. healthcare environments. And provide also supporting improvement in clinical decision-making and patient care.