Abstract
Today, data is stored in databases in a wide range of fields. Consequently, a significant amount of data has been generated over time. The value of these massive data sets lies in the knowledge and relationships that exist between them. The data mining process involves the analysis of data to obtain knowledge and facts that assist decision makers in making the best judgments possible. The use of data mining techniques in medicine has been around for generations, which is one of the fields where data mining is most important. This study aims to establish is to determine the encroachment of feature selection algorithms on classifier accuracy (model). There are 25 features in the dataset used to diagnose Chronic Kidney Disease (CKD), was utilized in this study to see how features selection techniques affect classifier accuracy. Here, Wrapper features selection evaluators are used to select those features that improve classification accuracy. We used classical Naive Bayes and J48 classifiers in this study and when Naive Bayes is used, accuracy increases from 95% to 99.5% as a classifier with a wrapper features selection evaluator. It is not a significant degree of accuracy, when j48 classifiers are used with wrapper features selection evaluators.