Proactive Diagnostics: Machine Learning-Driven Early Detection of Chronic Kidney Disease

Chronic Kidney Disease (CKD) is a growing global health challenge, contributing to 10–15 % of the worldwide population's morbidity and causing 1.7 million deaths annually. Early and accurate detection is crucial for preventing severe complications like hypertension, anemia, and neurological disorders. This study presents a machine learning approach for predicting CKD using clinical data from the UCI CKD dataset, with an emphasis on four major classification algorithms: Random Forest, Decision Tree, SVM, and Naive Bayes. The study involves data preparation by handling missing values using a filtering technique, and for feature selection, a Random Forest Classifier along with Mutual Information and Pearson correlation were applied. Among the models evaluated, the Random Forest Classifier led by an accuracy of 100%, while Decision Tree, Naive Bayes, and SVM achieved accuracy of 0.975, 0.75, and 0.9916, respectively. Random Forest's handling missing data, bias reduction, and complex interactions made it effective for CKD predictions.