摘要
随着信息网络的不断发展,现实生活中的网络往往是由多种类型的节点、链接组成的异质信息网络,传统的针对单一类型的同质网络表示学习方法不能直接应用于异质网络。近年来,越来越多的研究者提出针对异质信息网络的表示学习方法。本文首先介绍了异质信息网络表示学习的发展历程,然后将现有的异质信息网络表示学习方法进行分类,并对各个类别下的代表性算法进行系统性的介绍,简单阐述了异质信息网络表示学习的应用场景,最后给出了该领域未来的发展方向。
With the continuous development of information network,the network in real life is often heterogeneous information network composed of various types of nodes and links.The traditional representation learning method for a single type of homogeneous network can not be directly applied to heterogeneous network.In recent years,more and more researchers have proposed representation learning methods for heterogeneous information networks.This paper first introduces the development process of representation learning in heterogeneous information networks,then classifies the existing representation learning methods of heterogeneous information networks,systematically introduces the representative algorithms in each category,briefly describes the application scenarios of representation learning in heterogeneous information networks,and finally gives the future development direction of this field.
作者
张蝶依
尹立杰
ZHANG Die-yi;YIN Li-jie(School of College of Information Engineering,Hebei GEO University,Hebei Shijiazhuang 050031,China)
出处
《新一代信息技术》
2021年第8期24-29,共6页
New Generation of Information Technology
基金
河北省科技厅软科学研究计划(项目编号:17456001D)。
关键词
网络表示学习
异质信息网络
网络分析
Network representation learning
Heterogeneous information network
Network analysis
