Computer Engineering and Applications ›› 2019, Vol. 55 ›› Issue (1): 35-41.DOI: 10.3778/j.issn.1002-8331.1801-0255

Previous Articles     Next Articles

Identification of Node Influence Based on Improved k-shell Algorithm

ZHU Xiaoxia, HU Xiaoxue   

  1. School of Economics and Management, Yanshan University, Qinhuangdao, Hebei 066004, China
  • Online:2019-01-01 Published:2019-01-07



  1. 燕山大学 经济管理学院,河北 秦皇岛 066004

Abstract: Nodes that have greater influence in complex networks play an important role in controlling rumors propagation, optimizing resource allocation, spreading information efficiently, and advertising accurately. In view of the current many methods in identifying the node’s influence have certain limitation, this paper based on the k-shell algorithm defines the concept of weighted degree, and puts forward the Modified k-shell(MKS) algorithm, shorted for MKS algorithm by measuring the potential importance of edges and considering the different contributions of neighbors. This algorithm considers the nodes’own features, location features and local features. Through implementing this algorithm on the representative Zachary karate club network and comparing with other typical methods, it is found that this algorithm improves the coarse division of k-shell algorithm, and its result is more reasonable.

Key words: complex networks, k-shell, weighted degree, influence identification

摘要: 在复杂网络中具有较大影响力的节点在控制谣言传播、优化资源分配、高效传播信息、精确投放广告等方面发挥着重要作用。鉴于当前众多方法在识别节点的不同影响力时存在一定局限性,因此在k-shell方法的基础上,通过度量边的潜在重要性,考虑邻居节点的差异贡献性,从而定义了节点的加权度概念,并提出了MKS(Modified k-shell)算法,该算法综合考虑了节点的本身、位置以及局部属性。通过在具有代表性的Zachary空手道俱乐部网络上进行实现,并和其他典型方法进行比较分析,发现该算法改进了k-shell方法的粗粒化划分,其结果更加合理。

关键词: 复杂网络, k-shell, 加权度, 影响力识别