基于三角模体的团发现算法.PDFVIP

第41卷 第1期 南京理工大学学报 Vol.41 No.1 2017年2月 Journal of Nanjing University of Science and Technology Feb.2017 基于三角模体的社团发现算法 孙圣波,朱保平,杨晓光 (南京理工大学 计算机科学与工程学院,江苏 南京210094) 摘摇 要:为了提高社团发现算法的效率,提出了一种基于三角模体和期望极大的社团结构发现 (Community structure discovery based on triangular motifs and expectation鄄maximization,CSDTME) 模型的社团发现算法。 CSDTME模型采用三角模体对网络进行表示,考虑了节点的混合隶属度 及社团间的链接关系,用期望极大算法计算模型涉及的参数,采用全三角模体和两边三角模体 作为计算对象,通过减少计算对象来提高算法的效率,根据参数结果可得到节点的社团隶属度 及社团间的链接关系。 实验结果表明:在保证社团发现能力的同时,该算法能够提高社团发现 的效率。 关键词:三角模体;社团发现;期望极大算法;混合隶属度;链接关系 中图分类号:TP311摇 摇 文章编号:1005-9830(2017)01-0035-06 DOI:10.14177/ ki.32-1397n.2017.41.01.005 Community discovery algorithm based on triangular motifs Sun Shengbo,Zhu Baoping,Yang Xiaoguang (School of Computer Science and Engineering,Nanjing University of Science and Technology, Nanjing210094,China) Abstract:Inordertoimprovetheefficiency of community detectionalgorithm,thispaperproposesa community structure discovery based on the triangular motifs and expectation鄄maximization models of a community discovery algorithm.The model based on the triangle motif representsthe network, considering the links between nodes and mixed membership between communities.The expectation maximization algorithm is used to solve the parameters of the model,triangle motif and bilateral triangular norm body as an object of calculation by reducing the calculation object to improve the efficiency of the algorithm. The results are obtained according to the parameters of node membership linksandassociationsbetween communities.Theexperimentalresultsshow thattheal鄄 gorithm can improve the efficiency of the community discovery and ensure the capacity of the 摇 收稿日期:2016-08-0