Hello All,
I am looking for a good method for community detection that incorporates edge weights and if possible allows nodes to have overlapping membership in groups. Maybe it's wishful thinking but has anyone implemented or thinking about implementing the method in Farkas et al (2007) Weighted network modules (
http://www.iop.org/EJ/abstract/1367-2630/9/6/180)? Alternatively, does it make sense that the the 'leading.eigenvector.community' should be generally applicable to and could be modified for weighted networks? I realized that this won't allow for overlap in group membership.
Thanks for any suggestions and help
Colin