基于度与H指数扩展的复杂网络节点排序方法

在复杂网络中常用的识别节点影响力的中心性指标有介数中心性、度中心性、紧密中心性、H指数中心性和K-shell中心性等,这些指标在识别有影响力的节点时存在一定的局限性.本文在H指数中心性与度中心性的基础上提出了局部DH指数中心性指标来识别网络中有影响力的节点,该指标考虑了节点自身的度与H指数以及邻居节点的H指数.通过SIR传播模型以及单调函数(M)两种方法评价了各中心性方法识别网络中节点影响力的有效性.实验结果分析表明,在一些网络中该指标较一些常用的中心性方法能够更有效地识别网络中节点的影响力.

A method for sorting nodes in complex networks based on degree and H index expansion

The importance of identifying influential nodes in complex networks has attracted more and more attentions. Commonly used centrality indexes to identify the influence of nodes include betweenness centrality, degree centrality, closeness centrality, H-index centrality, K-shell centrality and so on. These indexes have some limitations in identifying influential nodes. On the basis of H-index centrality and degree centrality, a local DH index centrality index is proposed in this paper to identify influential nodes in the networks. This index takes into account the degree and H-index of the node itself and the H-index of neighboring nodes as well. The proposed index can evaluate the effectiveness of each centrality method to identify the influence of nodes in the networks by using SIR propagation model and monotone function (M). Our experimental results show that this proposed index can identify the influential nodes in some networks more effectively than commonly used centrality index.