A Comparative Analysis of the Ensemble Method for Liver Disease Prediction

Abstract

Early diagnosis of liver disease is very important in order to save human lives and take appropriate measure to control the disease. In several fields, especially in the field of medical science, the ensemble method was successfully applied. This research work uses different ensemble methods to investigate the early detection of liver disease. The selected dataset for this analysis is made up of attributes such as total bilirubin, direct bilirubin, age, sex, total protein, albumin, and globulin ratio. This research mainly aims at measuring and comparing the efficiency of different ensemble methods. AdaBoost, LogitBoost, BeggRep, BeggJ48 and Random Forest are the ensemble method used in this research. The study shows that LogitBoost is the most accurate model than other ensemble approaches.