Classification of Imbalanced Dataset using Generative Adversarial Nets
Emirhan Özmen, Fuat Çoğun, Fatih Altıparmak · 2020
One of the most encountered problems in the training of artificial neural networks is imbalanced datasets. It is common to apply oversampling algorithms to overcome the adverse affects of imbalanced datasets in classification performance. In this study, a new oversampling method based on Generative Adversarial Nets (GAN) and Edited K-Nearest Neighbor (KNN) is proposed. It is observed that the proposed method increases the classification performance more than that of the oversampling techniques based on the conventional SMOTE algorithm.