PERSONALIZED RECOMMENDATION METHOD THAT INDIVIDUALLY CONSIDERS EACH USER'S RECOMMENDATION ACCURACY AND DIVERSITY

摘要

A customized recommendation method that individually considers a user's recommendation accuracy and diversity according to an embodiment of the present invention includes the steps of: preprocessing data required for GAN learning; learning a first GAN for preference and a second GAN for diversity using the preprocessed data; generating recommendation information, which is information about individual users' preferences and diversity of all products, by using the user's feedback collection information and the learned first and second GANs; recommending top N products using the recommendation information; and reflecting the user's feedback on the N recommended products to the user's feedback collection information.