Title: Extension of the stochastic block model to handle networks with weighted nodes with application to EEG data
Authors: Yousri Slaoui - University of Poitiers (France) [presenting]
Abstract: The focus is on the analysis of weighted networks, finite graphs where each edge is associated with a weight representing the intensity of its strength. We introduce an extension of the binary stochastic block model (SBM), called binomial stochastic block model (bSBM). This question is motivated by the study of co-citation networks in a text mining context where the data is represented by a graph. The nodes are words and each edge joining two words is weighted by the number of documents included in the corpus simultaneously citing that pair of words. We develop an inference method based on the variational expectation maximization (VEM) algorithm and another based on the Bayesian variational expectation maximization (BVEM) algorithm to estimate the parameters of the proposed model as well as to classify the words of the network. Then we adopt a method based on the maximization of an integrated classification likelihood (ICL) criterion to select the optimal model and the number of clusters. Applications to real data are adopted to show the effectiveness of both methods as well as to compare them. Finally, we develop a multi-attribute SBM to handle networks with weights associated with nodes. We motivate this method with an application that aims at developing a tool to help specify different cognitive processes performed by the brain during writing preparation.