SharedData»Santo Fortunatos Community Detection Benchmarks

Santo Fortunatos Community Detection Benchmarks

Attach:santo_benchmark.tar

Santo writes: Dear Communities' Friends,

you may have seen the recent paper

http://arxiv.org/abs/0805.4770

in which Andrea Lancichinetti, Filippo Radicchi and myself introduce a new class of benchmark graphs, that build upon Girvan-Newman benchmark by adding some realistic features of real networks, i.e. the heterogeneous distributions of degree and community size. The benchmark poses a much harder test to algorithms of community detection, and I hope that the community will use it in the future instead of the Girvan-Newman benchmark (...) If you do not want to waste time to create the code to produce the graphs, don't worry. You can download it from

http://santo.fortunato.googlepages.com/inthepress2

along with a little guide, that explains everything.