Explainable Artificial Intelligence (XAI) for Heart Disease Prediction

This research focuses on building an Explainable Artificial Intelligence (XAI) system that helps make machine learning–based heart disease prediction more transparent and trustworthy for medical professionals. Although modern AI models can achieve impressive accuracy in healthcare, many of them operate as “black boxes,” making it difficult for clinicians to understand why a particular prediction is made. This lack of clarity often limits their acceptance and use in real clinical settings. To bridge this gap, the study presents a web-based diagnostic platform that combines powerful machine learning models—such as Random Forest and XGBoost—with well-established explainability techniques, namely SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations). The models are trained using the widely used UCI Heart Disease Dataset, which includes important clinical features like age, blood pressure, cholesterol levels, and chest pain type. The platform is designed to give doctors clear, real-time explanations for every prediction. It allows them to see not only how the model behaves overall but also why it produced a specific result for an individual patient. Through an intuitive Streamlit-based interface, clinicians can enter patient details, view heart disease risk predictions, and explore interactive visualizations such as SHAP summary plots, feature dependency graphs, and LIME-based local explanations. In addition to measuring standard performance metrics like accuracy, F1-score, and AUC-ROC, the system is evaluated for its practical usefulness and interpretability through simulated clinical cases and feedback from healthcare professionals. Early testing shows that adding explainability significantly improves transparency, builds clinician confidence, and supports better-informed medical decision-making. Overall, the study highlights that explainability is not just an optional feature but a critical requirement for the responsible and effective use of AI in sensitive medical applications. Future improvements could include extending the system to predict multiple diseases, integrating it with Electronic Health Records (EHRs), supporting mobile health platforms, and deploying it on the cloud to enable scalable and collaborative clinical use.