GWO BASED OPTIMISTIC FEATURE SELECTION FOR PREDICTION OF ADVANCED LIVER FIBROSIS

The present populace of India is 1,362,255,678 as of January, 2019, in view of the most recent United Nations estimates [1]. Worldwide examinations calculate that there are 8.7 million individuals living with chronic Hepatitis in India. Chronic Hepatitis disease represents 12-32 per cent of liver malignant growth and 10-20 per cent of cirrhosis cases in India. The vast majority with constant Hepatitis B or C are ignorant of diseases and are at genuine danger of creating cirrhosis or liver malignant growth.Machine learning fits a few procedures superior to other traditional methods like Biopsy. This paper evaluate various machine learning methods to predict advanced liver fibrosis by using patient’s blood report to build up the optimization and classification models. The METAVIR score is a device used to assess the seriousness of fibrosis seen on a liver biopsy test from a human who has chronic hepatitis. Based on the METAVIR score [2,3,4] chronic hepatitis divided into three parts, first one is classified as mild stage, second one is moderate stage and third one is advanced stage of fibrosis. Grey Wolf Optimization, Random Forest Classifier and Decision tree procedure models forpropelled fibrosis chance expectation were produced. ROC curve and confusion matrix was evaluated to compare the accuracy of proposed methods.