基于交叉度的有向网络中心节点识别算法研究  

Identification algorithms of center node in directed complex networks based on cross degree

作　　者：周立欣[1] 刘臣[1] 霍良安 王育清[1] 

机构地区：[1]上海理工大学管理学院,上海200093

出　　处：《计算机应用研究》2016年第11期3299-3302,3306,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(71401107;71303157);国家教育部人文社科基金资助项目(12YJZH126;14YJCZH71);上海市自然科学基金资助项目(13ZR1458200)

摘　　要：利用K-核分解的方法识别中心节点,被认为在复杂网络重要节点发现中具有特殊的优势。但K-核分解法在有向网络中只能简单地利用节点的出度、入度或者两者之和进行分解,不能区分两者的差异。针对这一问题,将有向网络中出度与入度的概念相结合,提出交叉度(cross degree)的概念;并利用交叉度提出识别有向网络中心节点的C-核分解法。该算法在无向网络中退化为K-核分解法。通过仿真实验和分析,发现该方法既保留K-核方法准确有效的优势,同时还具有较好的区分度,能够较好地识别有向网络中的重要节点。The K-shell decomposition algorithm is taken as an efficient method to identify influential node in complex net- works. Nevertheless, K-shell method cannot be used in directed network naturally, as it only simply used in-degree, out-de- gree or the sum of them to run the algorithm. This paper proposed a new measure of local importance of nodes in directed net- works based on out-degree and in-degree, called cross-degree. Then, it presented a novel algorithm to identify the center nodes in directed network using decomposition method like K-shell, which called C-shell. The algorithm degenerated to K- shell method in undirected network. Simulation results of epidemic spreading process on real network by using SIR model show that the C-shell decomposition algorithm can identify core nodes in directed networks defectively.

关 键 词：有向网络 中心节点 交叉度 c-核分解法 区分度 

分 类 号：TP399[自动化与计算机技术—计算机应用技术] TP393[自动化与计算机技术—计算机科学与技术]