Early Detection of Chronic Kidney Disease in Diabetic Patients using Explainable AI

Chronic kidney disease – is an increasing stage that substantially contributes to global morbidity and mortality. Early detection is important for slowing disease advancement and enhancing patient outcomes. However, conventional diagnostic methods often fail to provide early alerts due to the mild and progressive onset of CKD. This work utilizes machine learning methods to improve early detection accuracy while maintaining model clarity through Explainable AI (XAI). A detailed dataset was pre-processed, incorporating missing value analysis to enhance data integrity. Machine learning models, including Linear Discriminant Analysis (LDA), Random Forest (RF), Support Vector Machine (SVM), and Logistic Regression (LR), were used and examined for their predictive effectiveness. Logistic Regression was chosen for additional analysis using SHapley Additive exPlanations (SHAP) due to its linear structure, which enhances a more logical interpretation of attribute influences. SHAP was utilized to detect key biomarkers influencing CKD advancement, thereby improving model clarity and aiding clinical judgment. This work aims to improve CKD detection using ML and AI techniques, promising model interpretability and reliability in medical fields.