Heart Disease Prediction Using AI Models: A Comparative Study on the Sulianova Dataset

This study presents a comparative evaluation of five machine learning classifiers—Logistic Regression, Ridge Classifier, Linear SVC, ExtraTrees Classifier, and XGBoost—for predicting heart disease using the Sulianova dataset. After extensive data preprocessing, including outlier removal, feature engineering, and normalization, XGBoost achieved the highest performance (accuracy: 74.02%, AUC: 0.81). The research highlights the effectiveness of ensemble models and the importance of clinically informed features in cardiovascular disease prediction.