Community detection in networks is one of the most popular topics of modernnetwork science. Communities, or clusters, are usually groups of verticeshaving higher probability of being connected to each other than to members ofother groups, though other patterns are possible. Identifying communities is anill-defined problem. There are no universal protocols on the fundamentalingredients, like the definition of community itself, nor on other crucialissues, like the validation of algorithms and the comparison of theirperformances. This has generated a number of confusions and misconceptions,which undermine the progress in the field. We offer a guided tour through themain aspects of the problem. We also point out strengths and weaknesses ofpopular methods, and give directions to their use.
translated by 谷歌翻译