An Efficient Early Identification of Chronic Kidney Disease using Machine Learning Techniques

Chronic Kidney Disease (CKD) affects approximately 10% of the global population and is often asymptomatic in its early stages, delaying diagnosis and treatment. This study focuses on early identification of CKD using machine learning (ML) techniques, leveraging the UC Irvine CKD dataset containing 400 patient records with 25 clinical variables. To ensure data quality, the K Nearest Neighbor Imputer Algorithm was employed for effective handling of missing values, alongside consistent data normalization. Feature selection methods, including Recursive Feature Elimination with Cross-Validation (RFECV) and Univariate Feature Selection (UFS), were used to enhance model robustness and accuracy. Among the six ML classifiers evaluated, K Nearest Neighbor, Random Forest, Support Vector Machine, XGBoost, AdaBoost, and Gaussian Naive Bayes, the Random Forest classifier outperformed others with a remarkable accuracy of 99.5%. This study demonstrates the transformative potential of ML in healthcare by enabling precise and early CKD prediction, paving the way for timely clinical interventions and improved patient outcomes.