Infinite sparse block model with node textual information using the innovation process

The roles and the interactions among the roles which nodes play has a significant impact on the structure of the network. In order to estimate this underlying latent structure, its number and composition, flexible treatment of the structural uncertainty and efficient use of available information becomes a key issue. We take a Bayesian nonparametric approach, jointly modeling sparse network, node textual information and potentially unbounded number of components to handle the aforementioned task. We show using synthetic dataset that our model successfully learns the underlying structure and outperforms previous method.