Matrix Means for Signed and Multilayer Graph Clustering

Hong Kong

We study suitable matrix functions to merge information that comes from different kinds of interactions encoded in multilayer graphs, and its effects in cluster identification. We consider a family of matrix functions, known as power means, and show that different means identify clusters under different settings of the stochastic block model. For instance, we show that a limit case identifies clusters if at least one layer is informative and remaining layers are potentially just noise.