Node-importance identification in complex networks via neighbors average degree

Centrality metrics of complex networks are often used to study vulnerability of networks. Thus, node-importance research are with theoretical significance and wide-spread applications. Based on neighbors average degree, this paper presents a novel approach, which combines degree of a node and the average degree of its neighbors to measure node importance. By cascading failures experiments on real-world networks and synthetic networks, the proposed approach, which can cause more damage on the network topology, measures the node importance better then three existing centrality metrics. The results also indicates that node with lower neighbors average degree compared with peers with the same degree but higher neighbors average degree, is more important for network topology.