Hybrid Fuzzy Feature Augmentation And XG Boost Learning For Chronic Kidney Disease Detection

Abstract

Chronic kidney disease (CKD) is a progressive health disorder, and timely diagnosis is crucial to prevent serious complications. Machine learning techniques are extensively used for disease prediction; however, traditional models face limitations such as uncertainty in clinical data and class imbalance. To overcome these challenges, this study develops a Hybrid Fuzzy–XGBoost classification model for CKD prediction. In the proposed method, numerical clinical features are transformed into fuzzy linguistic levels like “low”, “medium”, and “high” using triangular membership functions, which better represent the uncertainty in clinical measurements. Additionally, a fuzzy sample weighting method is employed to improve the accuracy of CKD case identification. The expanded dataset is then used to train the XGBoost classification algorithm. The model's performance was assessed using Accuracy, Precision, Recall, F1-Score, and ROC–AUC. Experimental results demonstrated that the proposed Hybrid Fuzzy–XG Boost model outperformed traditional classification techniques, achieving 98% Accuracy, 100% Recall, and 0.984 AUC. These findings underscore the high predictive capability of this model for CKD diagnosis.