Chronic kidney disease (CKD) is a progressive renal disorder that affects approximately 10–15% of the global population and represents a significant contributor to morbidity and mortality worldwide. Early and accurate diagnosis of CKD is essential to enable timely clinical intervention and prevent progression to end-stage renal disease. This study presents a comparative analysis of two machine learning algorithms, namely Random Forest (RF) and Support Vector Machine (SVM), for the classification of CKD using clinical and laboratory data. The dataset was obtained from the UCI Machine Learning Repository, comprising 400 instances with 24 features and 1 binary target variable. Data preprocessing steps included missing value imputation, categorical encoding, outlier detection using the Interquartile Range method, and feature normalization. Both algorithms were implemented using their default hyperparameters to ensure a fair comparison without optimization bias. The dataset was split into 80% training (320 samples) and 20% testing (80 samples) using stratified sampling. Model performance was evaluated using five metrics: accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC-ROC). The experimental results demonstrated that Random Forest achieved perfect classification performance with 100% across all evaluation metrics, while SVM attained 98.75% accuracy, 100% precision, 97.87% recall, 98.92% F1-score, and 100% AUC-ROC. The superior performance of Random Forest is attributed to its ensemble learning mechanism that aggregates predictions from multiple decision trees, effectively reducing overfitting and enhancing model robustness. These findings suggest that Random Forest is a highly reliable algorithm for CKD classification and holds significant potential as a clinical decision support tool for early disease detection.
Copyrights © 2025