一种基于密度峰值聚类的社区发现算法

ommunity detection is an important technique, which is used to discoverthe cluster structureon networks. The density clustering algorithm "Clustering by fast search and find of density peaks"(DP) is extended for network data. This paper proposes a new local density and dissimilarity measure for nodes, so that the DP algorithm can be applied to the community discovery on networks. The extended algorithm not only inherits the advantages of the original algorithm which can find the class structure of any shape, but also it can quickly and effectively process network data. Finally, the experimental analysis on several real network data shows the effectiveness of the new algorithm.