“…Most-popular data-level approaches include extensions of the SMOTE algorithm into a multi-class setting [22,23,24], strategies using feature selection [25,26], and alternative methods for instance generation by using Mahalanobis distance [27,28]. Algorithm-level solutions include decision tree adaptations [29], cost-sensitive matrix learning [30], and ensemble solutions utilizing Bagging [31,32] and Boosting [5,33].…”