Optimized Machine Learning Framework for Chronic Kidney Disease Prediction with Near-Perfect Accuracy

A serious worldwide health issue, chronic kidney disease (CKD) needs to be identified early for treatment to be successful and patient outcomes to be better.In this work, we introduce a thorough machine learning-based framework for CKD prediction based on laboratory and clinical characteristics.Different feature selection and train-test split strategies were used to evaluate several classifiers, such as logistic regression, naive Bayes, support vector machines, decision trees, and random forests.To determine the most important characteristics influencing an accurate diagnosis, feature ranking techniques like ReliefF and OneR were used.Simple baseline classifiers such as the majority class's ZeroR predictor, which consistently predicted CKD and only attained roughly 62% accuracy, were also taken into consideration in order to verify the efficacy of the model.According to experimental results, Decision Tree and Random Forest models outperformed the baseline method by achieving 100% accuracy, sensitivity, specificity, and ROC values.These findings demonstrate how well-tuned machine learning models can support automated, dependable, and extremely accurate CKD prediction systems.With useful ramifications for clinical decision support and early disease detection, the findings illustrate the importance of feature optimization and model selection in developing reliable healthcare diagnostic tools.