Clustering algorithm ideas

[smitalm]

Hello,

i will be implementing several clustering algorithms for my master thesis, and i decided for gephi as visualization tool.

So far i only saw 2 clustering algorithms - MCL experimental and Chinese Whispers implemented for gephi. Are there any other algorithms currently in development?

Are there any specific clustering algorithms that are missing from gephi, and you would like to see them implemented?

[admin]

Hi,

We only provide these 2 algorithms, but be free to implement your own ones...

[evilefi]

Hi there.

So the community can wish for graph clustering algorithms that you consider to implement? Sounds great!

The following would be interesting from my point of view:

1)
"Finding communities in linear time: a physics approach" by Fang Wu and Bernardo Huberman
http://www.hpl.hp.com/research/idl/pape ... linear.pdf
(maybe this could help: http://jung.sourceforge.net/doc/api/edu ... terer.html)

2)
"SCAN: A Structural Clustering Algorithm for Networks" by Xiaowei Xu, Nurcan Yuruk, Zhidan Feng and Thomas A. J. Schweiger
http://www.ualr.edu/nxyuruk/publications/kdd07.pdf
(where detected hubs and outliers could consitute single-node clusters in Gephi with a matching labeling of the cluster)

3)
Finally it would be interesting to have some attribute-aware clustering, s. a. in:
"Graph Clustering Based on Structural/Attribute Similarities" by Yang Zhou, Hong Cheng and Jeffrey Xu Yu
http://www1.se.cuhk.edu.hk/~hcheng/summ ... 09-175.pdf

Thanks for your willingness to contribute!

[smitalm]

Hello, thank you for your suggestions.

I liked SCAN algorithm and decided to implement it. I also implemented heuristics for interactive parameter selection (mentioned in SCAN paper).

Next algorithm i will implement will be this one:
Link communities http://barabasilab.neu.edu/projects/linkcommunities/.