[Top][All Lists]

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

Re: [igraph] Dear Prof. :I have a question about your "R igraph package"

From: Tamas Nepusz
Subject: Re: [igraph] Dear Prof. :I have a question about your "R igraph package".
Date: Thu, 19 May 2011 16:20:38 +0200
User-agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv: Gecko/20110424 Lightning/1.0b2 Thunderbird/3.1.10

Dear Qingju,

steps is usually set to 3, 4 or 5 in walktrap.community; it controls the
length of random walks the algorithm makes on the graph in order to
determine the affinity of vertex pairs to each other. The default value in
the original implementation of the authors
(http://www-rp.lip6.fr/~latapy/PP/walktrap.html) is 4; in general, it should
be long enough so that the random walks convey enough information about the
mesoscale structure of the network, but not too long. The original walktrap
paper contains more information about the topic:


As for the "steps" parameter of community.to.membership: the walktrap method
creates a dendrogram, and you must cut that dendrogram after a given number
of merges in order to obtain a partition. "steps" specifies the number of
merges to take in the dendrogram before stopping. Typically you should cut
the dendrogram at the level which gives the highest modularity score; this
can be achieved with the following code snippet in R:

wt <- walktrap.community(G, modularity=TRUE)
wmemb <- community.to.membership(G, wt$merges,
memberships$`Walktrap` <- wmemb$membership

See http://igraph.wikidot.com/community-detection-in-r#toc4 for more
information about how to use the community detection algorithms.


On 05/19/2011 05:04 AM, qingju588 wrote:
> Dear Prof.,
>        I am a postgraduate majoring in bioinformatics inShanghai,China. I 
> used the R igraph package (R igraph_0.5.1) to mine modules in networks. But I 
> do not how to set a optimal parameter for "steps" in "walktrap.community" and 
> "community.to.membership" functions?
>                    Thank you very much!
>                                                             Sincerely yours
>                                                            Qingju Jiao
> _______________________________________________
> igraph-help mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/igraph-help

reply via email to

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