关键词: 社团结构，模块密度，核k-means算法

Abstract: Discovery community structure is fundamental for uncovering the links between structure and function in complex networks. In this context, recently, Li et al. recently proposed modularity density objective function for community detecting called the U function and gave the equivalence between modularity density objective function and the kernel k-means by using a kernel matrix In this paper, based on this ectuivalence, we used the kernel matrix to optimize the modularity density and developed a new kernel k-means algorithm. Experimental results indicate that the new algorithms arc efficient at finding community structures in complex networks.

Key words: Community structures, Modularity density, Kernel k-means