Performance Evaluation of Advanced Machine Learning Models for Chronic Kidney Disease Classification and Prediction

Abstract

Chronic Kidney Disease (CKD) is a major global health concern, with increasing prevalence and significant impacts on patient morbidity and mortality. Early and accurate prediction and classification of CKD can greatly assist in timely interventions, improving patient outcomes and resource allocation in healthcare. This paper presents an extensive analysis of nineteen machine-learning models for prediction and classification in CKD. The performance of each model is assessed using confusion matrix metrics such as Accuracy, Precision, Recall, and F1-Score, offering a thorough evaluation of their predictive abilities. Moreover, ROC curves and AUC scores are employed to evaluate and compare the models' ability to distinguish between classes. Results indicate that certain models are more effective in handling the unique characteristics of CKD data, providing insights into how algorithm selection affects predictive performance. The Random Forest classifier outperformed the others, achieving an accuracy of 0.99, along with precision, recall, and F1-score values all at 0.99. This paper underscores the potential of machine learning advancements and predictive modeling in creating innovative solutions, particularly for improving prediction accuracy in kidney disease and similar fields. Also offers a comprehensive resource for selecting suitable machine learning models in CKD applications and highlights areas for future research in CKD-related predictive analytics.