Visualization and classification of graph-structured data: the case of the Enron dataset

Graph-structured networks are often used to represent relationships between persons in organizations or communities. In this paper we investigate the problem of learning a latent space representation of the data in which proximity in the latent space increases the likelihood of a social tie between the nodes. In addition, this latent space representation can be used to classify these data into homogeneous groups in order to identify, for instance, marginal communities of persons. We propose a Bayesian way to select both dimension of the latent space and number of groups. We apply our approach to the Enron dataset and we show interesting representation and clustering of individuals.