A Machine Learning Approach to Chronic Kidney Disease

Abstract

Chronic kidney disease (CKD) is a worldwide health problem that causes significant morbidity and mortality, as well as the onset of other illnesses. People frequently miss CKD because there are no obvious symptoms in the early stages. Early detection of CKD allows patients to receive timely treatment to slow the disease's progression. Machine learning models can successfully assist doctors in achieving this goal due to their rapid and precise identification capabilities. In this paper, we present a machine learning framework for CKD diagnosis. The CKD data set was retrieved from the University of California, Irvine's machine learning repository (UCI). Due to this, it will determine whether a patient has CKD and, if so, whether or not additional medications need to be taken. Models were developed using six machine learning techniques: gradient boosting, logistic regression, adaBoost, random forest, and decision trees. The most accurate machine learning model was random forest. We proposed an integrated model that combines logistic regression and random forest using perceptron, best accuracy, by examining the errors produced by the existing models. We therefore hypothesised that this methodology might be applicable to clinical data for disease diagnosis that is more complex.