一种对分划分的复杂网络社团检测方法

在许多领域,例如社会科学,技术科学及生物科学,复杂网络中的社团发现是一项重要任务。这些社团结构暗含着系统功能方面的信息并用来帮助人们理解网络的功能及增长机制。谱分优化了由李等人最近提出的一种用来评估和发现社团的模块密度函数。提出了一种对分算法,该算法使用模块密度矩阵的主特征向量迭代来检测网络社团结构。在一个经典的计算机产生的随机网络中检验了算法。当社团结构变地模糊时,实验结果显示这种新的算法在发现复杂网络社团上是有效的。

Detecting of communities in complex networks with two partitioning approach

Discovery of community structures in complex network is a fundamental task in many fields, for instrance, social science, technology science and biology science. These community structures imply information about system function, and it can be used to help people understand the function of network and its growth mechanism. We optimize modularity density function to spectral questions, and then pro- pose a two partitioning algorithm which uses the leading eigenvectors of the modularity density matrix to split a network into communities. The algorithm is illustrated and compared with spectral clustering based on modularity （Q） using a classic computer generated network. The experimental results show that the proposed approach is effective, particularly when community structure is obscure.