Disclosed are a generative adversarial network (GAN) based classification system using labeled data and unlabeled data, and a classification method thereof. The GAN-based classification system can be trained by using a labeled data set and an unbalanced data set such as an unlabeled data set by using missing data generated by using a GAN. The GAN-based classification system comprises a generator (100), a discriminator (200), an actor (400), a weight function part (500), and a reward part (600).;COPYRIGHT KIPO 2020
展开▼