In recent years, complex network theory with advances in the understanding of the highly interconnected nature of various social, biological, and communication systems has gained much attention[1, 2, 3, 4, 5, 6, 7, 8]. Transferring information, trust, ideas, diseases and influences between any two nodes is the key function of complex networks[9, 10, 11]. The information can spread rapidly to a large number of nodes begin with an influential node. Hence, evaluating the influence of the nodes is a significant issue in complex networks [12], such as in the control of the disease and rumor dynamics[13, 14, 15], research on public opinion[16, 17], and creating new marketing tools [18, 19, 20, 21, 22, 23, 24, 25, 26, 27]. Many measurements of node centrality have been used commonly such as Degree centrality (DC)[28], Betweenness centrality (BC)[28, 29, 30], Closeness centrality (CC)[28] and so on. The DC method is very simple but of little relevance, since the measure does not take into consideration the global structure of the network. BC and CC are global metrics which can better identify influential nodes, but they are difficult to apply in largescale networks due to their computational complexity. Another limitation of CC is the lack of applicability to networks with disconnected components: two nodes that belong to different components but do not have a finite distance between them. Several spectral centrality measures are also available, such as semi-local centrality(SLC)[31], eigenvector centrality (EC)[32], PageRank (PR)[33], and LeaderRank (LR)[34]. In SLC the topo logical connections among the neighbors are neglected, only the number of the nearest and the next nearest neighbors of a node is taken into account[35]. EC can not be applied to asymmetric networks in which some positions are unchosen[32]. PR, as well as LR, only has effect in directed networks, it will degenerate to DC in undirected networks[31].