Bidirectional dynamic recommendation system based on generative adversarial network

• Main Authors: XIA BIN, HAN YUE, FANG XIAOYU, CAO CHENHAN
• Format: Patent
• Language: Chinese; English
• Published: 23.02.2021
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses a bidirectional dynamic recommendation system based on a generative adversarial network. The system employs two generative adversarial networks, wherein one trains a generatorin the generative adversarial network and learns a subject selection object through a universal data set, and the other one generative adversarial network trains a generator in the generative adversarial network, learns the preference of the object selection subject by using the universal data set, and finally recommends the subject and the object in the reciprocal recommendation system by using the converged generative adversarial network obtained by training. Multiple recommendation options are flexibly generated, so that the preference of the user can be captured more quickly. 本发明公开了一种基于生成对抗网络的双向动态推荐系统，采用了两个生成对抗网络，其中一个训练生成对抗网络中的生成器并利用通用数据集学习主体选择客体，另一个训练生成对抗网络中的生成器并利用通用数据集学习客体选择主体的偏好，最后用训练得到的已经收敛的生成对抗网络分别对互惠推荐系统中的主体以及客体分别进行推荐，本发明能够根据用户相关性以及用户选择同时迭代，灵活地生成多个推荐选择，对用户的偏爱性捕捉更加敏捷。