Assessment of Chronic Kidney Disease using Classification Algorithms

Abstract

Kidney Disease (CKD) implies the condition of kidney risk which may even get worse by time and by referring the factors. If it continues to get worse Dialysis is done and worst-case scenario it may lead to kidney failure (End-Stage Renal Disease). Detection of CKD in an early stage could help in sorting out the complications and damage. In the previous work classification used are SVM and Naïve Bayes, it resulted that the execution time took by Naïve Bayes is minimal compared to SVM, incorrect instances are less for SVM that results in less classification performance of Naïve Bayes, because of slight accuracy difference. It can be rectified by taking a smaller number of attributes. Naïve Bayes is a probabilistic classifier a simple computation by applying Bayes Theorem with a conditional independence assumption. The work mainly results in increasing diagnostic accuracy and decrease diagnosis time, that is the main aim. An attempt is made to develop a model evaluating CKD data gathered from a particular set of people. From the model data, identification can be done. This work has engrossed on developing a system based on classification methods: SVM, Naïve Bayes, KNN.