为有效地检测复杂网络中的社团结构,优化模块密度函数,展示模块密度函数怎样被优化框定到谱分聚类问题,提出一种谱分算法,进一步对该算法进行时间复杂度分析。在一个经典的真实世界网络中检验该算法,并与基于模块密度的直接核方法及基于模块函数的谱分方法做比较。特别地,当网络中社团结构变得模糊时,实验结果显示,该谱分算法在发现复杂网络社团上是有效的。
To detect community structure in complex networks, modularity density function is optimized. By optimizing process, how to optimize the D function can be reformulated as a spectral relaxation problem and a spectral clustering algorithm is proposed. Furthermore, the time complexity of.algorithm is analyzed. The approach is illustrated and compared with direct kernel approach based on modularity density and spectral clustering based on modularity by using a classic real world networks. Experimental results show the significance of the proposed approach, particularly, in the cases when community structure is obscure.