Chronic Kidney Disease Prognosis through Machine Learning Approaches

Chronic Kidney Disease (CKD) is a condition that damages the kidneys and affects their ability to remove waste and maintain fluid balance in the body. It can lead to complications such as high blood pressure, anemia (low red blood cell count), bone disorders, poor nutrition, acid base imbalances, and nerve problems. Detecting CKD early and accurately is crucial for preventing further health decline. Its high prognosis for developing into end-stage renal disease calls for reliable prediction tools to enhance early detection and therapy. This research focuses on improving CKD prediction using machine learning (ML) techniques. The study applies advanced methods like Boosting, Random Forest, SVM, k-Nearest Neighbors (KNN) and a Voting Classifier, combining algorithms like AdaBoost, Support Vector Machine (SVM), and KNN. Models were evaluated using metrics such as accuracy, precision, recall, and F1 score to identify the best model for early prediction. The dataset for this study, sourced from the UCI Repository, contains 400 entries and 21 attributes. It was analyzed using various classification algorithms and by comparing various machine learning approaches, the goal is to find the best predictive models to help CKD early and improve care for affected patients. The results showed impressive accuracy, with the Random Forest model achieving 98.75% and Gradient Boosting reaching 98.25%. This research highlights the potential of ML to improve early CKD detection and support better healthcare strategies.