Comparative Analysis of Machine Learning Models for Predicting Heart Disease

Heart disease remains to be a foremost universal health issue, highlighting the need for precise and effective predictive tools to support early diagnosis and timely intervention. This work presents an $\mathbf{A I} / \mathbf{M L}$ prediction system designed to detect heart diseases by comparing and evaluating multiple machine learning algorithms. The proposed system utilizes a comprehensive dataset that includes a variety of patient attributes such as demographic, clinical, and lifestyle factors, gathered from multiple healthcare sources. Through preprocessing and feature engineering, the dataset is organized for training and testing the machine learning models. To achieve accurate predictions, this work appraises the efficiency of some prevalent ML algorithms, including decision trees, random forests, logistic regression, Bayes, support vector machines, and k-nearest neighbors. A thorough comparative analysis is conducted, assessing the algorithms based on significant scientific metrics such as recall, accuracy, F1-score, precision and AUC-ROC, to determine their suitability for heart disease prediction.