Role Detection: Network Partitioning and Optimal Model of the Lumped Markov Chain
Nowadays, complex networks are present in many fields (social science, chemistry, biology, …) as they allow to model systems with interacting agents. In many cases, the number of interacting agents is large (from hundreds to millions of nodes). In order to get information about the functionality of the underlying system, we are interested in studying the structure of the network. One way to do that is by partitioning the network. In this paper, we present a method to detect a partition of the network such that the dynamics of a random walker on the lumped network is a good model of the dynamics of a random walker in the original network.
We acknowledge support from the Belgian Programme of Interuniversity Attraction Poles and an Action de Recherche Concertée (ARC) of the French Community of Belgium.
- 1.Aynaud T, Guillaume JL (2011) Multi-step community detection and hierarchical time segmentation in evolving networks. In: Proceedings of the 5th SNA-KDD workshop Google Scholar
- 4.Cason T (2012) Node-to-node similarity measures and role extraction in networks. PhD thesis, Université catholique de Louvain, Belgium Google Scholar
- 5.Cooper K, Barahona M (2010) Role-based similarity in directed networks. E-print. arXiv:1012.2726
- 9.Lambiotte R, Delvenne J-C, Barahona M (2009) Laplacian dynamics and multiscale modular structure in networks. arXiv:0812.1770