Chronic Kidney Disease Detection Augmented with Hybrid Explainable AI

This study proposes an explainable machine learning (ML)-based framework for early prediction of chronic kidney disease (CKD), addressing the critical challenge of model interpretability in AI-driven healthcare applications. Various ML algorithms (random forest, support vector machines, gradient boosting, and logistic regression) are optimized using feature selection techniques (i.e., RFE and SelectKBest) to enhance diagnostic accuracy. Ensemble models achieve 100% accuracy, demonstrating the effectiveness of ML in CKD detection. To ensure transparency, explainable AI (XAI) techniques such as SHAP, fuzzy rule-based systems, and decision trees are applied to identify key biomarkers (e.g., hemoglobin, serum creatinine, and specific gravity), making predictions clinically interpretable. The integration of fuzzy logic further aligns model decisions with medical reasoning, enhancing clinician trust. This research bridges the gap between AI predictions and clinical decision-making, contributing to the development of transparent, data-driven clinical decision support systems for early CKD detection, personalized treatment, and improved patient outcomes.