A machine Learning Methodology for Chronic Kidney Disease

Abstract

Most of the people in the world are suffering a lot with Chronic Kidney Disease (CKD) widely considered to be a global issue. People with heart disease, hypertension, diabetes or who have suffering from CKD family they are at risk.This is related to improved mortality and high risk of a few other conditions, including the raise of heart disease and human services. The adverse outcome of CKD can be prevented with proper early detection and management. Machine Learning (ML) aids to analyze tens of thousands of data points and produce outcomes, provides regular risk scores, precise resource allocation, and has many other applications healthcare. In this paper some machine learning techniques are used and compared namely Support Vector Machine (SVM), Naïve Bayes (NB), NB-TREE, and Iterative Random Forest (IRF). Among all these four algorithms IRF got the best accuracy of 99.4%. Here we are using IRF which provide the stable decision paths with high order interactions. Hence IRF could be used for any data to get good accuracy.