网络顶点表示学习方法  

Approaches on network vertex embedding

作　　者：周晓旭 刘迎风 付英男 朱仁煜 高明 ZHOU Xiaoxu;LIU Yingfeng;FU Yingnan;ZHU Renyu;GAO Ming(School of Data Science and Engineering,East China Normal University,Shanghai 200062,China;Shanghai Municipal Big Data Center,Shanghai 200072,China)

机构地区：[1]华东师范大学数据科学与工程学院,上海200062 [2]上海市大数据中心,上海200072

出　　处：《华东师范大学学报（自然科学版）》2020年第5期83-94,共12页Journal of East China Normal University(Natural Science)

基　　金：国家重点研发计划(2016YFB1000905);国家自然科学基金(U1911203,U1811264,61877018,61672234,61672384);中央高校基本科研业务费专项;上海市科技兴农推广项目(T20170303);上海市核心数学与实践重点实验室资助项目(18dz2271000)。

摘　　要：网络是一种常用的数据结构,在社交、通信和生物等领域广泛存在,如何对网络顶点进行表示是学术界和工业界广泛关注的难点问题之一.网络顶点表示学习旨在将顶点映射到一个低维的向量空间,并且能够保留网络中顶点间的拓扑结构.本文在分析网络顶点表示学习的动机与挑战的基础上,对目前网络顶点表示学习的主流方法进行了详细分析与比较,主要包括基于矩阵分解、基于随机游走和基于深度学习的方法,最后介绍了衡量网络顶点表示性能的方法.Network is a commonly used data structure, which is widely applied in social network,communication and biological fields. Thus, how to represent network vertices is one of the difficult problems that is widely concerned in academia and industry. Network vertex representation aims at learning to map each vertex into a vector in a low-dimensional space, and simultaneously preserving the topology structure between vertices in the network. Based on the analysis of the motivation and challenges of network vertex representation, this paper analyzes and compares the mainstream methods of network vertex representation in detail, including matrix decomposition, random walk and deep learning based approaches, and finally introduces the methods to measure the performance of network vertex representation.

关 键 词：网络嵌入 随机游走 矩阵分解 深度神经网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]