Robust chronic kidney disease prediction via synthetic data validation and interpretable ensemble learning

Chronic Kidney Disease (CKD) requires early and accurate diagnosis for effective clinical management. This study proposes an interpretable ensemble learning framework integrating Conditional Tabular Generative Adversarial Network (CTGAN)-based synthetic data augmentation, Recursive Feature Elimination with Cross-Validation (RFECV), and SHapley Additive exPlanations (SHAP) for CKD prediction. Four ensemble classifiers, Random Forest, XGBoost, LightGBM, and CatBoost, were evaluated on original and augmented datasets. The augmented models achieved up to 99.17% test accuracy with high precision, recall, and F1-score. Statistical, predictive, and visual evaluations confirmed the quality of the synthetic data. SHAP-based interpretability provided clinically meaningful insights, supporting reliable and explainable CKD prediction.