[R] Undirected to Directed Graph ( Clustering Based on Link Strength)

Sur Nathan surendar.swaminathan at gmail.com
Thu Mar 5 20:25:34 CET 2009


Hello All,
 I went through lot of searching for papers to do clustering on
Co-Authroship.I ran SNA metrics on my Undirected
Graph(betweenness,Closeness,Eigen Vector) using R.
 
After running the sna Metrics I found important nodes and did clustering on
those nodes based on Binding strength .This clustering is done through if
they are connected.I would like to do clustering based on Weight of the
link.
 
Example lets say 5 nodes have published 1 paper together and out of which 3
nodes in the 5 nodes have seperately published 10 papers.The link strength
between the 3 nodes would be greater than the 5 nodes combined together and
3 nodes should be clustered.
 
My understanding is Adjacency matrix or Dissimilarity Matrix would not let
me cluster the ihighly interconnected node.
 
When I went through nabble R Archive and I found many message but I do not
think there was any message in regard to this.
 
I did find one Article by I have attached in this mail.

Co-authorship networks in the digital library research community
 
Xiaoming Liu ,*, Johan Bollen , Michael L. Nelson , Herbert Van de Sompel 
 
They have done exactly what I am looking for.They have done bottom up
hierarchical Clustering.
 
Is there anyway I can do this in R.
Help on this would be great.

Thanks a lot
Nathan http://www.nabble.com/file/p22358708/Clustering%2BTechnique.pdf
Clustering+Technique.pdf 
-- 
View this message in context: http://www.nabble.com/Undirected-to-Directed-Graph-%28-Clustering-Based-on-Link-Strength%29-tp22358708p22358708.html
Sent from the R help mailing list archive at Nabble.com.




More information about the R-help mailing list