Neighbor vector centrality of complex networks based on neighbors degree distribution

We introduce a novel centrality metric, the neighbor vector centrality. It is a measurement of node importance with respect to the degree distribution of the node neighbors. This centrality is explored in the context of several networks. We use attack vulnerability simulation to compared our approach with three standard centrality approaches. While for real-world network our method outperforms the other three metrics, for synthetic networks it shows a slightly weak properties but still a good measure overall. There is no significant correlation of our method with network size, average degree or assortativity. In summary, neighbor vector centrality presents a novel measurement of node importance, which has a better performance to reduce dynamics of real-world complex networks.