Main partitioning methods

Historically, the search for communities was a matter for sociologists, as can be seen in the SW problem. Then, considering that the problem falls within the domain of graphs, mathematical computation, essentially matrix computation, takes over, followed by algorithmic computation on computers, which favors links. A third approach, which considers the problem to be one of data analysis, calls on the computational methods of this field.

Most of the work was first carried out on unipartite graphs and then, more recently, on bipartite graphs, with a few rare forays into multipartite and oriented graphs. Weighting is often introduced to direct the detection algorithm. Several explanations can be found for this focus on unipartite graphs....