***** To join INSNA, visit http://www.insna.org ***** Many thanks to everyone who responded with links to publications, programs, and websites for me to check out.  It's very much appreciated!

One follow up question for the group related to community identification.  What is your opinion on community algorithms that create overlapping lists (1 person can belong to many communities) vs. those that create mutually exclusive lists (1 person can only belong to 1 community)?  

When doing analysis that has problems with fuzzy boundaries I can see the benefit of algorithms that create mutually exclusive lists, however, when dealing with human networks it seems that the mutually exclusiveness of something like Newman Girvan would prevent us from getting a true picture of the communities that exist within a network since people tend to belong to multiple communities.

All the best!
~ Steve

On Fri, Jan 1, 2010 at 4:31 AM, Mason Alexander Porter <[log in to unmask]> wrote:
*****  To join INSNA, visit http://www.insna.org  *****

My first community detection e-mail of 2010... :)

There are a couple of issues to consider based on the question asked.  One is how to construct communities in the first place, and then there's your cohesiveness measure.  One thing that I'd strongly suggest is to do your cohesiveness calculations using multiple methods of defining communities to see how much things change.  The results might not be robust with respect to your definition (they probably won't be), so you're going to need to be careful.  The G-N betweenness-based method isn't going to be your best choice (as has been pointed out), though the issue is significantly broader than an implied one (from earlier in this thread) about the application context of the network you're studying.  It tends to have trouble with highly-clustered networks, for instance.

In terms of which method to use, I can make further comments (I tend to be a bit opinionated about this particular area), but aside from any private correspondence with further details, let me just add a couple of hopefully helpful articles to the mix:

1. I'll advertise my own survey/review article on community structure in networks, which I think is a good place to get started on the topic (obviously, I'm very biased about this) and see the layout of the land:

       http://people.maths.ox.ac.uk/~porterm/papers/comnotices.pdf


2. The article in (1) is a survey and as you're actually doing the calculations, I think you'll also want to go to the following encyclopedic review by Fortunato:

       http://arxiv.org/abs/0906.0612

(you might want to check if the published version of this is now out; it will come out pretty imminently)



I think my article is pretty user-friendly, so usually I recommend starting there and then getting into the more technical articles.

As I mentioned, I'm happy to provide more info (and several of the previous e-mails mentioned lots of good stuff), so let me know.

-----
Mason

----------------------------------------------------------------------------
 Mason A. Porter
 University Lecturer (and Tutorial Fellow, Somerville College)
 Oxford Centre for Industrial and Applied Mathematics
 Mathematical Institute, University of Oxford

 Homepage: http://www.maths.ox.ac.uk/~porterm, IM: tepid451
 Blog: http://masonporter.blogspot.com/
----------------------------------------------------------------------------
 "I don't know. Maybe the knowledge of asymptotic analysis will lead
 to less starvation among African children?"

       --- Me, in an early draft of a grant proposal when asked to address
               how the project will help with the socio-economic
               development of third-world countries
----------------------------------------------------------------------------

_____________________________________________________________________
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.insna.org). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.

_____________________________________________________________________ SOCNET is a service of INSNA, the professional association for social network researchers (http://www.insna.org). To unsubscribe, send an email message to [log in to unmask] containing the line UNSUBSCRIBE SOCNET in the body of the message.