Print

Print


*****  To join INSNA, visit http://www.insna.org  *****

As regards overlaps, there's also the following paper which seemingly  
proposes a quite simple and elegant approach consisting in applying  
traditional partitioning algorithms on link graphs, rather than node  
graphs.

Evans, TS, Lambiotte, R, Line graphs, link partitions, and overlapping  
communities, Phys Rev E, 2009, Vol 80.
http://arxiv.org/pdf/0903.2181

As a result, link partitions yield relatively straightforward  
decompositions of graphs into (possibly) overlapping node subsets.

On Jan 1, 2010, at 4:57 PM, Steve Eichert wrote:
> (...)
> 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)?
> (...)


(Admittedly -- and it's plausibly a broader issue touching other  
approaches as well -- this kind of method could probably be more or  
less sociologically relevant in order to appraise social circles/ 
overlapping 'communities')

Camille

On Jan 4, 2010, at 8:34 PM, Kathleen Carley wrote:
> *****  To join INSNA, visit http://www.insna.org  *****
>
> FOG (available in ORA http://www.casos.cs.cmu.edu/projects/ora/) and  
> see
> the Social Network article - also let's you have an overlapping  
> structure.
>
> Davis, George & Carley, Kathleen. (2008). Clearing the FOG:
> Understanding Interstitial Relationships with Fuzzy, Overlapping  
> Groups.
> Social Networks, 30, 201-212.
> http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6VD1-4S2F5N0-1&_user=525223&_rdoc=1&_fmt=&_orig=search&_sort=d&_docanchor=&view=c&_acct=C000026389&_version=1&_urlVersion=0&_userid=525223&md5=a3fb5517648bb87f9adae0cfe8a7ecba
>
> James Moody wrote:
>> *****  To join INSNA, visit http://www.insna.org  *****
>>
>> Hi -
>>
>> The node-connectivity / cohesive blocking approach that Doug White  
>> and I
>> have proposed similarly allows overlapping structure, so it's not a  
>> complete
>> "blind spot"; and does so in a way where the overlap structure can be
>> formally explicated.  See:
>> http://www2.asanet.org/journals/ASRFeb03MoodyWhite.pdf .  The  
>> advantage of
>> node-connectivity for understanding social cohesion is the clear link
>> between diffusion-through-multiple paths and higher-orders of
>> k-connectivity.
>>
>> Best,
>> Jim
>>
>>
>> -----Original Message-----
>> From: Social Networks Discussion Forum  
>> [mailto:[log in to unmask]] On
>> Behalf Of Balazs Vedres
>> Sent: Sunday, January 03, 2010 6:02 AM
>> To: [log in to unmask]
>> Subject: Re: community identification algorithms: KliqueFinder
>>
>> *****  To join INSNA, visit http://www.insna.org  *****
>>
>> There seems to be a blindspot in sociology for overlapping groups.  
>> We used
>> the Clique Percolation Method:
>>
>> http://angel.elte.hu/cfinder/
>>
>> that yielded interesting theoretical benefits to specify the  
>> significance of
>> structural folding (the overlapping of groups) for entrepreneurial
>> recombination and generative tension. Business groups do tend to  
>> overlap a
>> lot, and overlapping groups are better performing, but more unstable.
>> Needless to say, you need a historical dataset to understand what  
>> cohesion
>> means (as Simmel, Moreno, and the early theoretical school of  
>> network ties
>> clearly saw).
>>
>> http://www.personal.ceu.hu/staff/Balazs_Vedres/papers/vedres.stark.folds.pdf
>>
>> This article appears in the January issue of AJS.
>>
>> _____________________________________________________________________
>> 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.
>>
>
> _____________________________________________________________________
> 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.

--
cams (cnrs-ehess) / crea (cnrs-ep) / cress (u. surrey)
http://camille.roth.free.fr  -  gsm +33 (0)6 8034 0870

_____________________________________________________________________
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.