Use of Machine Learning for Early Detection of Chronic Kidney Disease (CKD)

Abstract

Chronic Kidney Disease (CKD) poses a major global health burden due to its gradual onset and often silent progression. Traditional diagnostic methods, based on a limited set of laboratory markers, may delay detection until significant kidney damage has occurred. Machine learning (ML) offers promise for early detection by analyzing complex, multi-dimensional patient data to identify subtle patterns indicating early kidney dysfunction. In this study, we evaluate several ML classifiers — including Logistic Regression, Support Vector Machine (SVM), Random Forest (RF), Gradient Boosting (GB), and k-Nearest Neighbors (KNN) — on publicly available clinical datasets. After preprocessing, feature normalization, inconsistency handling and class balancing, models are trained and evaluated. The experimental results show that ensemble-based methods outperform individual classifiers, with Random Forest achieving the highest accuracy (≈ 98.6%) and robustness to noisy clinical data. These results underscore the potential of ML-based diagnostic tools to support early CKD screening, enabling timely medical intervention and improved patient outcomes.