基于混合的K核影响力的标签传播算法

Community Detection Base on mixed K-Influence

社会网络中近年来的一个研究热点就是社团发现,在众多的社团发现算法中,将标签传播算法应用到社团发现中是非常快速的一种算法,得益于其接近线性的时间复杂度,该算法能够处理大规模的社会网络。KLPA是一种基于k值影响力的改进标签传播算法,结合K值及局部影响力改进迭代及标签选择过程。实验证明,使用改进算法能够得到更好的社区划分结果。

Community detection has become a hotspot of social network analysis. Compared to other community detection al-gorithms, LPA is a very fast algorithm. Because of its nearly-linear time in detection, it can be used in large social network. KLPA is an proposed algorithms based on k-value. It combined K-Core decomposition and local influence to modify the process of iteration and label selection. Tests on classic and real social networks demonstrate that the proposed algorithm can do a better detection in social networks.