摘要
简要介绍复杂网络社区发现的常用方法,对比社区发现的几种主流方法。根据图谱分析法和特征值分析方法,提供一种基于主成分分析的方法来揭示社区结构,并通过实验数据对比证明发现结果的准确性,解释此种变换的理论基础,同时说明划分结果的正反社区结构与特征值的关系。最终达到从降维角度进行社区发现的目的。
We give a brief introduction to the commonly used methods of complex community discovery, and compare several mainstream community discovery algorithms. Based on spectral analysis method and eigenvalue analysis method we provide a principal component analy- sis-based method to reveal community structure, and through the comparison of experimental data to prove the accuracy of the result, as well as explain the theoretical basis of such transformation. Meanwhile we illustrate the relationship between the positive and negative community structure of the partitioning result and the eigenvalue. Finally we achieve the purposes that to discover the community from perspective of di- mension reduction.
出处
《计算机应用与软件》
CSCD
2015年第9期261-263,268,共4页
Computer Applications and Software
关键词
复杂网络
社区
图谱分析法
Complex network Community Spectral analysis method
