From the data scientific point of view, network analysis at the macroscopic level (extraction of communities, cliques, and other structural blocks) is an example of unsupervised machine learning. The goal of unsupervised machine learning is to infer a network’s hidden structure in the absence of “labels”: node and edge attributes (except, perhaps, the edge weights).
The unearthed blocks suffer from two major interrelated problems:
In fact, if you knew the purpose or nature of a block, you would give it a name, and if you knew the name, you would guess what its purpose or nature is.
Selecting a good name for a block can be done in at least three ways.
You can use your intelligence: ...