A Novel approach to Handle Imbalanced Dataset in Machine Learning

Abstract

The world has seen an exponential rise in machine learning and artificial intelligence since the 1990s. We apply machine learning models to solve various real life problems like regression and classification. However, class imbalance is a very common issue faced for classification problems in machine learning. In this study, we propose new greedy resampling techniques to solve the problem of class imbalance. We shall also compare the results of these techniques with the Synthetic Minority Over-sampling Technique (SMOTE).