链接相似性的微博重叠社区发现算法.doc

链接相似性的微博重叠社区发现算法
于洪涛,崔瑞飞,黄瑞阳
(国家数字交换系统工程技术研究中心,郑州 450002)
E-mail:cuiruifei0815@
摘要:针对传统基于节点聚类的微博社区发现算法不能发现重叠社区且需要先验知识这一问题,本文从边聚类的角度出发,提出了一种基于链接相似性的微博重叠社区发现算法。首先将用户兴趣相似度矩阵映射为虚拟兴趣网并求该网络的链接相似度,然后结合微博用户的真实关注关系得到总的链接相似度。为了将链接相似度用于社区发现,推广了传统的Ward层次聚类算法,使之适用于具有相似性度量的任意对象,并将其用于社区发现。真实数据集上的实验表明,该算法不需要先验知识就能准确地发现微博中的重叠社区,%。
关键词: 虚拟兴趣网;关注网络;链接相似性;层次聚类;重叠社区
中图法分类号:TP393 文献标识码:A
A link-based similarity micro-blog munity detecting algorithm
YU Hong-tao, CUI Rui-fei, HUANG Rui-yang
(National Digital Switching System Engineering & Technological R&D Center, Zhengzhou 450002, China)
Abstract: It is impossible to detect munity with traditional node clustering algorithms and they need priori-knowledge. To solve the problem, this paper proposed a link-based similarity micro-blog munity detecting algorithm. Firstly, the paper mapped the interest similarity matrix to a work and sought its link similarity and then attained the total link similarity bing real attention relationship among users. To utilize link similarity munity detection, we generalize the Ward hierarchical clustering algorithm so that it is applicable to any object that has similarity measurement. And as an application we particularly employ this algorithm to munity. Experiments on real data sets show that the algorithm can detect micro-blog munity without priori-knowledge, with the accuracy of %.
Keywords: vir