Abstract: Imbalanced class distribution disrupts the training of a classifier, resulting in biases favoring majority classes. Data oversampling is a common strategy to tackle this issue. However, ...