XGBoost-Enhanced Prediction and Interpretation of Heart Disease Using SHAP Values

Heart disease represents a critical challenge to global public health, with its propensity to gravely impact human life and well-being. While current prediction models have achieved relative success, the heterogeneity and complexity inherent in medical data often result in diagnostic inaccuracies when applied practically. There is, therefore, a pressing need for the development of prediction models that are both more accurate and efficient. This article introduces a classification model for heart disease prediction, underpinned by the XGBoost algorithm and informed by authentic medical datasets from the Kaggle platform. To elucidate the predictive decisions of the model, SHapley Additive exPlanations (SHAP) values are utilized, which provide an interpretive framework for the output. The proposed model exhibits strong predictive capabilities and interpretability, presenting a valuable tool for medical practitioners in the diagnosis and management of heart disease.