I. Introduction
Imbalanced data classification, a prominent issue captivating the machine learning and data mining community, has emerged as one of the ten formidable challenges in contemporary research [1]. This challenge is particularly pronounced in binary classification scenarios where datasets exhibit a significant disparity in the number of samples belonging to each class, creating what is commonly referred to as an imbalanced dataset [2] [3]. Imbalance occurs when one class, designated as the minor class, comprises a substantially smaller percentage of samples compared to the other class, known as the major class. In such scenarios, the minor class is often characterized by a scarcity of instances, rendering it more challenging for machine learning algorithms to discern patterns and make accurate predictions. The implications of imbalanced data are far-reaching, influencing the effectiveness of machine learning models and hindering their ability to generalize accurately across diverse applications. This issue becomes particularly critical in scenarios where the minority class represents outcomes of significant interest, such as fraud detection or the identification of rare diseases. The skewed distribution of samples introduces complexities, as learning algorithms tend to be biased toward the majority class, leading to suboptimal performance in terms of sensitivity and specificity.