BAGGING BASED ENSEMBLE ANALYSIS IN HANDLING UNBALANCED DATA ON CLASSIFICATION MODELING

Abstract

The purpose of this study is to Identify the algorithm of each method of handling the unbalanced class based on bagging based on the literature review. This study uses a bagging based ensemble method such as UnderBagging, OverBagging, UnderOverBagging, SMOTEBagging, Roughly Balanced Bagging and the last one is the Bagging Ensemble Variation. The data used is coded from the UCI Repository with 16 data, eight of which have class categories with low imbalance problems, and the rest are categorized as high imbalance problems. The number of classes used in this study amounted to two classes. The class with a small number is made into the minority class and the rest is made up as the majority class. The result of this research is the bagging based method gives better results when compared to classical methods such as the classification tree.