[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[igraph] Problem in leading eigenvector community detection

From: Deepika Roy Chowdhury
Subject: [igraph] Problem in leading eigenvector community detection
Date: Mon, 19 Jul 2010 03:27:34 +0530


                                                  I am Deepika. I have recently started using the igraph library and i am finding it very useful. I am a Ph.D student in University of Delhi, India. I read the paper written by M.Newman - "Finding community structure in networks using the eigenvectors of matrices"  and decided to use the algorithm to find communities in networks. And therefore used igraph for it. In the igraph reference manual - the definition of the 'matrix merges' isn't clear to me and therefore I don't know what it is doing. Also when i use the code for a network having nodes = 297 , it gives the following warning : 
                                                         Warning: Maximum number of ARPACK iterations reached in file community.c, line 1176 

Have attached a file "i8f.c" consisting of 10 nodes. 
Used "igex6.c" (igraph leading eigenvector community detection program) to find communities in the above network. And the output is stored in "o8.txt" file

In the output, it prints the matrix merges first and don't know what each element of the matrix implies. I'd very much like to know what it is. 
Can you tell me what needs to be done to avoid the above warning for large networks.
I'd also like to know what should be the value of 'steps'  for large networks . Is 'steps = no. of nodes'  the only last best possibility?

Does this method give good results for large networks or should I also try the other methods(written in igraph) for community detection.

I would appreciate your help.


Attachment: o8.txt
Description: Text document

Attachment: i8f.c
Description: Text Data

Attachment: igex6.c
Description: Text Data

reply via email to

[Prev in Thread] Current Thread [Next in Thread]