We introduce a novel class of stochastic blockmodel for multilayer weighted networks that accounts for the presence of a global ambient noise that governs between-block interactions. In weighted multilayer noisy networks, a natural hierarchy is induced by the assumption of all but one block being under the `umbrella' of signal, while a single block is classified as exhibiting the same characteristics as interstitial interactions and subsumed into ambient noise. We use hierarchical variational inference to jointly detect block-structures and differentiate the communities' local signals from global noise. We propose a novel algorithm to find clusters in Gaussian weighted multilayer networks with a focus on these principles called Stochastic Block (with) Ambient Noise Model (SBANM).We apply this method to several different domains. We focus on the Philadelphia Neurodevelopmental Cohort to discover communities of subjects that form diagnostic categories relating psychopathological symptoms to psychosis.
View on arXiv