<![CDATA[Chronic Kidney Disease (CKD) is one of the major health problems that continues to rise and requires accurate early detection to prevent progression to end-stage renal failure. This study proposes a hybrid machine learning approach to automatically detect CKD by combining data balancing techniques, ensemble learning, and cross-validation. The dataset used was obtained from the Kaggle platform, consisting of 1,089 patient records, and was balanced using the Synthetic Minority Over-sampling Technique (SMOTE) to address class imbalance. Three boosting algorithms—Adaboost, XGBoost, and LightGBM—were used as base models and combined through a stacking approach with Logistic Regression as the meta-classifier. Evaluation was conducted using a 5-fold cross-validation scheme with accuracy, precision, recall, and F1-score as performance metrics. The results show that the stacking model achieved an average accuracy of 99.40%, outperforming individual models (LightGBM: 98.87%; Adaboost: 98.76%; XGBoost: 98.61%) and exceeding the performance of several previous studies. These findings indicate that the stacking approach, when combined with SMOTE and cross-validation, significantly enhances classification performance for CKD detection.]]>
Copyrights © 2025
