Review on Data Balancing Approaches for Skewed Data Classification

Addressing the skewed data set is a significant concern nowadays. In machine learning, using such data to learn distributions poses a significant challenge. This problem arises when there are significantly more samples of one class, also known as a majority class, than of the other classes. The more important class is known as a minority. As a result of its use in so many real-world applications, researchers’ interest in it is increasing. This issue of class imbalance has been studied using a variety of methods over the past few years, including methods such as data resampling, ensemble techniques, cost-sensitive analysis, hybrid methods, etc. With an emphasis on binary-class and multi-class problems, the goal of this study is to discuss the current state of the art on various strategies to comprehend the state of machine learning strategies for enhancing classification output and addressing class imbalance issues.