See Mark Newman's solution for bipartite networks.
http://www.pnas.org/content/98/2/404.full
Uzzi and Spiro apply this method in a study of Broadway Musical Teams.
http://www.kellogg.northwestern.edu/faculty/uzzi/ftp/uzzi's_research_papers/uzzi&spiroajs_smallworlds.pdf
On Wednesday, May 8, 2013 4:51:59 AM UTC-4, Tamás Nepusz wrote:
> Unfortunately I cannot just generate a random network as this will cause an underestimation of the PL and overestimation of the CC when the actual network is compared to the random network (due to the fact that unipartite projections of bipartite networks give cliques whenever e.g. inventors worked on the same patent).
How about generating a random bipartite network, projecting it into unipartite and then calculating the CC and the PL on the projection?
Cheers,
Tamas
_______________________________________________
igraph-help mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/igraph-help