A new approach for community detection in multilayer networks.

In this work, we propose a new approach for clustering multilayer and attributed networks, which captures the emergent behaviour of complex systems. The goal is to assign each network node (shared across network layers) to clusters, considering altogether the extra information carried by nodes and the connectivity patterns in each layer. This is a challenging task because one has to combine two types of information (Yang et al., 2013), while leveraging the extent to which topological and attribute information contribute to the network’s partition (Falih et al., 2018). We present an extension of the existing MultiTensor (MT) model recently developed by De Bacco et al. (2017), which performs an overlapping community detection task on multilayer networks by taking into account the interactions among the system’s constituents. Specifically, we describe MultiTensorCov (MTCov): this model considers both sources of information for uncovering groups of nodes that are structurally close but also share some common characteristics.