A Hybrid Approach to Heart Attack Risk Prediction Using Ensemble and Traditional ML Models

Abstract— Heart attacks are one of the most serious types of cardiovascular disease [18] (CVD), which is a leading cause of death worldwide. Early and accurate heart attack prediction can greatly enhance the resultes for patient by enabling timely intervention. This study assesses the likelihood of a heart attack using a machine learning-based predictive model that uses Random Forest, XGBoost, Logistic Regression, Support Vector Machine (SVM), and Naïve Bayes [11][12][13][14][15] based on patient data.The dataset is subjected to preprocessing methods such feature selection, data cleaning, and normalization in order to improve model performance. Key performance indicators, such as accuracy, precision, recall, F1-score [17], and AUC-ROC, are used to assess the model. XGBoost demonstrated the supremacy of ensemble approaches by having the greatest accuracy of 98.61% among the evaluated models, followed by Random Forest at 98.41%. The use of probabilistic methods in classification was highlighted by the somewhat lower accuracy of SVM and Logistic Regression, whereas Naïve Bayes offered a baseline performance. These results demonstrate the effectiveness of ensemble models like Random Forest and XGBoost in predicting heart attacks and the promise of machine learning in medical diagnostics. To further improve clinical application and predictive accuracy, future studies can concentrate on real-time data integration, deep learning developments, and model interpretability.