Diagnosis Heart Attack and Stroke Risk using Machine Learning Algorithms: Risk Classification and Explanation of Diagnostic Reasons with LIME

Abstract

A heart attack is a condition that can occur due to the blockage of blood vessels that carry blood pumped by the heart to other parts of the body, hereditary factors, or organ failure leading to the heart being deprived of oxygen. It is a life-threatening condition that may result in death. Stroke problems can also arise due to the inability to pump blood properly. In 2023, 33.4% of deaths in Turkey were caused by cardiovascular diseases. This study includes health findings, factors affecting heart health, and test results from 588 women and 1300 men. The analyses in this study were conducted using machine learning algorithms, and the results were used to make predictions categorized as high risk or low risk. Furthermore, the LIME (Local Interpretable Model-Agnostic Explanations) method was applied to explain the reasons for the predictions and the weights of the features influencing the predictions. Machine learning (ML) methods such as Random Forest, XGBoost, Decision Tree, Naive Bayes, Gradient Boosting, K-Nearest Neighbors, and Support Vector Machines were utilized in this study. Among the machine learning algorithms used, Random Forest achieved the highest success rate with 96.82%. Additionally, the LIME method was employed to provide examples explaining the decisions made by the algorithms. © 2025 IEEE.