Intelligent Heart Disease Prediction Using CatBoost Empowered with XAI

Abstract

Heart disease refers to a class of diseases involving the heart and blood vessels, including coronary artery disease, heart failure, and arrhythmias. Concerns associated with heart disease include increased risk factors like hypertension, smoking, high cholesterol, and a sedentary lifestyle. Early prediction and effective management are crucial to mitigate these risks and prevent complications. In the context of heart disease prediction, the CatBoost machine learning (ML) classifier was applied to a dataset divided into training and testing, resulting in a notable testing accuracy of 84.4%. The SHAP (SHapley Additive exPlanations) explainability technique was employed to illuminate the decision-making process of CatBoost predictions. Through computation and visualization of SHAP values, this approach enhances transparency, fairness, and interpretability, revealing insights into the pivotal features influencing heart disease prediction. The integration of the XAI (Explainable Artificial Intelligence) technique in this research seeks to deepen the knowledge of the complex dynamics involved in heart disease prediction.