Tang and Liu [42] cluster edges instead of nodes, which results in overlapping communities of nodes. Following up the latter work, Wang et al. [43] study the problem of discovering overlapping groups in social media, and they propose a coclustering framework based on users and tags.