Centralities in High Order Networks

Abstract

We propose a method for computing centralities based on shortest paths in time-varying, multilayer, and time-varying multilayer networks using MultiAspect Graphs (MAG). Thanks to the MAG abstraction, these high order networks are represented in a way that is isomorphic to a directed graph. We then show that well-known centrality algorithms can be adapted to the MAG environment in a straightforward manner. Moreover, we show that, by using this representation, pitfalls usually associated with spurious paths resulting from aggregation in time-varying and multilayer networks can be avoided.