Radial-Based Approach to Imbalanced Data Oversampling

摘要

The difficulty of the many practical decision problem lies in the nature of analyzed data. One of the most important real data characteristic is imbalance among examples from different classes. Despite more than two decades of research, imbalanced data classification is still one of the vital challenges to be addressed. The traditional classification algorithms display strongly biased performance on imbalanced datasets. One of the most popular way to deal with such a problem is to modify the learning set to decrease disproportion between objects from different classes using over- or undersampling approaches. In this work a novel preprocessing technique for imbalanced datasets is presented, which takes into consideration the mutual density class distribution. The proposed approach has been evaluated on the basis of the computer experiments carried out on the benchmark datasets. Their results seem to confirm the usefulness of the proposed concept in comparison to the state-of-art methods.