基于图的多视角聚类算法综述  

A Survey of Graph-based Multi-view Clustering Algorithms

作　　者：王春杰 何进荣 王文发[1] WANG Chunjie;HE Jinrong;WANG Wenfa(School of Mathematics and Computer Science,Yan'an University,Yan'an 716000)

机构地区：[1]延安大学数学与计算机科学学院,延安716000

出　　处：《计算机与数字工程》2022年第2期229-237,255,共10页Computer & Digital Engineering

基　　金：国家自然科学基金项目(编号:61902339,61763046);国家自然科学基金地区项目(编号:618660389)资助。

摘　　要：多视角聚类通过利用多视角之间的互补性和一致性信息来提高聚类的性能。近年来受到越来越多的关注。为了及时掌握目前基于图的多视角聚类算法的研究现状与最新技术,对大量的、最新的多视角图聚类进行调查、归纳整理、分类及总结。根据多视角聚类涉及的算法机制和数学原理,并进一步分为基于图、基于网络和基于谱的聚类方法。不仅详细介绍了每一类算法数学原理、算法模型,而且还举例说明了这些算法的应用。报告了基于图的多视角聚类的现状,最后总结了各类算法的优缺点,并指出了当下的挑战以及未来研究发展的方向。Multi-view clustering(MvC)can improve the performance of clustering by using the complementary and consistent information between multi-view. More and more attention has been paid to it in recent years. In order to grasp the current research status and the latest technology of Graph-based multi-view clustering algorithms,and a large number of the latest MvC surveys,induction,sorting,classification and summary. According to the Algorithm Mechanism and mathematical principle of MvC,and it is further divided into graph-based,network-based and spectral-based clustering methods. Not only the mathematical principles and models of each kind of algorithms are introduced in detail,but also the applications of these algorithms are illustrated with examples. This paper reports the status of graph-based MvC in which summarizes the advantages and disadvantages of various algorithms,and points out the current challenges and future research directions.

关 键 词：图聚类 网络聚类 谱聚类 多视角聚类 图构造 多维特征 图融合 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]