Chronic Kidney Disease Risk Prediction Using Machine Learning

Chronic kidney disease is one of the fast-growing global health issues that necessitates early detection for effective management and treatment. This research article examines the application of ML-based techniques to predict CKD, using clinical data for accurate classification. Three major algorithms are tested here for CKD stage prediction: RF, SVM, and MLP. We also adopt ensemble methods, including hard and soft voting, to make the predictive performance even stronger. Our results show that while SVM has the highest accuracy at 96.25, the ensemble methods provide very robust solutions, with the hard voting yielding an accuracy of 86.33. This research shows the prospect of ML models and techniques of ensemble in improving the diagnosis of CKD, valuable for clinical decision-making and treatment of patients. Future work will concentrate on optimizing these models and integrating them into clinical applications for real-time predictions.