Message Boards Message Boards

[✓] What's the meaning of communities in Wolfram Language?

GROUPS:

I'm trying some clustering analyses, and the references show me a group of functions in the aspect of social network analysis. To visualize the networks, one can use the function CommunityGraphPlot[].

The examples given in http://reference.wolfram.com/language/ref/CommunityGraphPlot.html show that the function can work out the community of the points automatically, but I do not understand how it work out the community, and what is the algorithm of the community identifying. By now there is no available plain explanation for it.

Any help will be appreciated.

POSTED BY: Linzhi Li
Answer
3 months ago

Community detection in graphs is a large topic. You can read an introduction e.g. here:

There are multiple commonly used methods for community detection. Mathematica implements several of these, but the details of the algorithms are not documented. You can look at the Method option of FindGraphCommunities and take reasonable guesses at what general type of method each of the possible settings uses (e.g. I expected "Centrality" to use the Girvan-Newman method).

To get the details on each method, I suggest you contact Wolfram Support. Personally, I found them to be helpful on most occasions (though others have sometimes reported a refusal to disclose the algorithm details...)

As you can see, you're not the first to ask about this:

As I mentioned in an answer to the second linked question above, the IGraph/M package exposes several community detection methods implemented in igraph. The igraph documentation usually has references to the original papers that introduced these methods.

POSTED BY: Szabolcs Horvát
Answer
3 months ago

Thank you very much! Your answer is really helpful.

I am feeling regretful for my blindness to the existed resources in this forum, maybe I over relied on google and typed in key words that are not so precise.

POSTED BY: Linzhi Li
Answer
3 months ago

Group Abstract Group Abstract