标题:
一种基于效用的社区发现算法A Community Discovery Algorithm Based on Utility
作者:
杨德品, 周丽华, 程超, 龙克珍
关键字:
社会网络, 社区发现, 效用, K均值Social Network; Community Discovery; Utility; K-Means
期刊名称:
《Operations Research and Fuzziology》, Vol.4 No.1, 2014-02-21
摘要:
本文从图论思想出发，提出了一种基于效用的社区发现算法，该方法既考虑了社区成员联系的频繁度又考虑了联系的重要度。本文定义了效用的概念，通过效用来描述节点相似度，并实现了基于效用的社区发现(Community Discovery Based on Utility，简写为CDBU)算法，该算法有效地避免了传统的基于联系频繁度的社区发现方法忽略了联系重要度的弊端。最后，本文在真实数据集上进行了实验，验证了所提出算法的合理性和有效性。Based on the graph theory, this paper proposes a community discovery algorithm based on utility. The method considers not only the contact frequency but also the importance of the links between community members. The paper defines the concept of utility to describe nodes similarity and implements the CDBU algorithm. The algorithm proposed effectively avoids the abuse of the traditional community discovery method based on the contact frequency, which neglects the important degree of contact. Finally, on the real-world dataset, we verify the rationality and validity of the algorithm proposed in this paper.