Classification of Chronic Kidney Disease with Genetic Search Intersection Based Feature Selection Technique

Abstract

The objective of this study had reduced the identification time and to get best diagnosis model in Chronic Kidney Disease (CKD). The target was development of the strong and computationally proficient model for classification of CKD. This work had used four classification models like Naive Bayes, Multilayer Perceptron, OneR, Classification and Regression Tree (CART) to classify the CKD data set and compared the classification accuracy with all classifier. Similarly Feature Selection Technique (FST), ranking based namely Chi Squared AttributeEval (CSAE), One-R AttributeEval (ORAE) and Search based namely Genetic Search-J48 (GS-J48), Genetic Search-CART (GS-CART) have used. The contribution of this research work is to recognize and classify the CKD problem and propose a new FST namely Genetic Search Intersection Based Feature Selection Technique (GS-IBFST). The classifier offers higher accuracy with GS-IBFST compare to without FSTs and existing FSTs.