Optimization of Support Vector Machine Classifier Using Grey Wolf Optimization Algorithm for Chronic Kidney Disease Prediction

Abstract

The massive generation of medical data from smart health-care applications in recent years necessitates the development of big data classification strategies. Medical data classification can be used to visualize patterns in the data and detect the presence of the disease in medical data. We present an efficient support vector machine (SVM) hybridized with a grey wolf optimization (GWO) algorithm for chronic kidney disease (CKD) data classification in this work. Initially, infinite feature selection (IFS) algorithm is used to select the best features from a set of available features. The dataset’s selected features are processed and fed into the GWO optimized SVM algorithm. The proposed CKD classification strategy has been simulated in MATLAB. CKD dataset from UCI machine learning repository is utilized for testing the developed strategy. The performance of the proposed CKD classification strategy is examined by accuracy and root mean square error (RMSE) values. According to the investigational findings, the proposed CKD classification system achieved accuracy and RMSE value of 97.58% and 0.1581, respectively, for classifying subjects into the CKD and non-CKD categories. The performance of GWO optimized SVM algorithm is outstanding, according to the experimental observations.