Multi-layer academic network community discovery method and system based on generative adversarial network model

• Main Authors: SUN QINGYUN, ZHU SHIJIE, JI CHENG, LI JIANXIN, FU XINGCHENG, DONG XIANGYU
• Format: Patent
• Language: Chinese; English
• Published: 29.05.2020
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

According to a multi-layer academic network community discovery method and system based on the generative adversarial network model, embedded representation of a multi-layer network is learned based on a GAN model, and the multi-layer academic network is constructed to learn node embedding representation by using a generative adversarial model; a generator generates an intra-layer node pair and aninter-layer node pair as pseudo samples, and a discriminator judges whether the data is real data distribution or not; the generator and the discriminator are iteratively updated to carry out adversarial learning. According to the method, a step of discovering a community based on a K-means clustering method is used to realize processing of scholar information of a network source, discover deep information of a relation network which can be provided by a multi-layer network structure, and enable an algorithm and a system of the method to have higher robustness. 本发明实现了一套基于生成对抗网络模型的多层学术网络社区发现方法及其系统，基于GAN模型学习多层网络的嵌入表示，通