Stochastic blockmodels: Graph representations and Community Detection

Partitioning a graph and its efficient representation appear to have an increasing importance, due to its implications in complex networks i.e. social networks. One important model that is utilized is called Stochastic blockmodels [link] where attributes of connections generates certain classification (blocks) that leads to a reduced representation or inherent partions. A recent work by Newman generalizes this model to real networks to detect communities [link], by taking into account the variation in the vertex degree.