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


Dear Socnet mailing list members.



Another community detection algorithm: VOS Clustering (Waltman and van Eck):

is now implemented in Pajek 3.05.

Thanks to Ludo Waltman and Nees Jan van Eck for providing details of the algorithm.

It is a Multi-Level Coarsening + Multi-Level Refinement algorithm as well,  

speed is approximatelly the same as in Louvain method - 

therefore it can be applied to large networks as well.




Implementation of Louvain community detection algorithm was further speeded up.

Additionally some new parameters were provided:

- number of random restarts - start the optimization several times and select the best result

- maximun number of iterations in each restart - for huge networks this might be useful.





The communities obtained by VOS Clustering are similar but not exactly the same

as the ones obtained by the Louvain method since another quality function is optimized

(VOS quality instead of modularity). For some comparisons of results obtained

by Louvain Method and VOS Clustering see:






Entering parameters for several procedures (e.g. random network generators) was simplified:

instead of one input box for each parameter, now only one input box with all parameters

that must be entered appears.




You are welcome to check the history and download Pajek 3.05 at:

or (mirror)




With best wishes.

Andrej Mrvar





_____________________________________________________________________ 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.