Measuring similarity for clarifying layer difference in multiplex ad hoc duplex information networks

Abstract In multiplex networks, each layer may represent different interactions or the same interaction over different time periods. Presently all centralities method may fail to detect the change among different layers (totally M layers). As the minimum unit of a multiplex network is duplex network (M = 2), we can clarify layer difference via duplex network. In a duplex network, the layer similarity LSim is defined for measuring similarity between layers, via node similarity of two layers, and then the layer difference is described by the similarity. The methodology can be extended to multiplex network by repeats of duplex networks. Two information networks and two extending empirical cases are investigated and verified.