考虑对象属性信息的复杂网络社团结构发现算法

针对社团结构发现算法仅考虑对象间相互关系的密集程度,忽视对象间属性特征差异的不足,提出考虑属性信息的复杂网络社团结构发现算法.算法引入属性特征相似度、基于属性特征相似度的有权网络、内聚度3个核心概念,迭代选取使内聚度指标上升最快的合并操作,自底向上实现社团聚集.由于考虑了属性信息,算法输出的社团结构具有更高准确度,更具应用价值.

Detecting Community Structure in Complex Network Based on Attribute Information of Objects

In order to improve the community structure detecting algorithms which focus on relationship concentration within a community but ignore the attribute difference among objects, the algorithm Detecting Community Structure in Complex Network Based on Attribute Information of Objects is proposed.After bringing in three core concepts named attribute similarity,weighted network based on attribute similarity and inner aggregation extent,the algorithm repeatedly merges pairs of communities that increase the inner aggregation extent at best,and assembles the community structure from bottom up.The analysis shows that the algorithm outputs better result with denser relationship and higher attribute similarity in communities,which is useful in most real applications.