A concept of higher order neighborhood in complex networks, introduced previously (PRE textbf{73}, 046101, (2006)), is systematically explored to investigate larger scale structures in complex networks. The basic idea is to consider each higher order neighborhood as a network in itself, represented by a corresponding adjacency matrix. Usual network indices are then used to evaluate the properties of each neighborhood. Results for a large number of typical networks are presented and discussed. Further, the information from all neighborhoods is condensed in a single neighborhood matrix, which can be explored for visualizing the neighborhood structure. On the basis of such representation, a distance is introduced to compare, in a quantitative way, how far apart networks are in the space of neighborhood matrices. The distance depends both on the network topology and the adopted node numbering. Given a pair of networks, a Monte Carlo algorithm is developed to find the best numbering for one of them, holding fixed the numbering of the second network, obtaining a projection of the first one onto the pattern of the other. The minimal value found for the distance reflects differences in the neighborhood structures of the two networks that arise from distinct topologies. Examples are worked out allowing for a quantitative comparison for distances among a set of distinct networks.
Download