Representative community divisions of networks

Abstract Methods for detecting community structure in networks typically aim to identify a single best partition of network nodes into communities, often by optimizing some objective function, but real-world applications there may be many competitive partitions with scores close the global optimum and one can obtain more informative picture examining representative set such high-scoring than looking at just optimum. However, difficult interpret since its size easily run hundreds or thousands partitions. In this paper we present method analyzing large sets dividing them groups similar then identifying an archetypal as each group. The resulting provides succinct, interpretable summary form variety any network. We demonstrate on range example networks.