<![CDATA[Chronic Kidney Disease (CKD) is an escalating global health concern that demands faster and more accurate diagnostic solutions than traditional laboratory-based assessments. This study evaluates and compares the performance of three machine learning algorithms Random Forest (RF), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) for early CKD prediction using the publicly available dataset from the UCI Machine Learning Repository. The research workflow includes comprehensive data preprocessing, handling missing values, addressing class imbalance using the Synthetic Minority Over-sampling Technique (SMOTE), and performing hyperparameter optimization with GridSearchCV combined with 5-fold cross-validation. The experimental results indicate that all evaluated models demonstrate strong predictive performance, with SVM achieving the best recall-oriented results, recording an accuracy of 0.93, a recall of 0.99, and an F1-score of 0.96. Feature importance analysis identifies Serum Creatinine, Glomerular Filtration Rate (GFR), Hemoglobin, Blood Pressure, and Age as the most influential clinical predictors of CKD. Overall, the findings demonstrate that integrating SMOTE-based imbalance handling with systematic hyperparameter tuning significantly enhances model robustness, while the SVM model provides high sensitivity, making it particularly suitable for early CKD screening and clinical decision-support applications.]]>
Copyrights © 2026
