4 ms·
Is there an easy way to compute the second smallest eigenvalue in question if the graph is large? There is a well-known theorem that quickly computes the small
by zzleeper 12y ago
Is there an easy way to compute the second smallest eigenvalue in question if the graph is large?
There is a well-known theorem that quickly computes the smallest eigenvalue if the graph is bipartite, but I'm not aware of any generalization (which may be quite useful for some results in econometrics).
- mathgenius 12y agoI recently found this paper: "Hearing the clusters in a graph: A distributed algorithm" http://arxiv.org/abs/0911.4729 http://arxiv.org/abs/0911.4729 Wish I had some data to apply it too :-)
- hugofirth 12y agoIt's been a while since I've done work in this area but you might want to lookup Lanzcos Algorithm [1]. [1]:http://en.m.wikipedia.org/wiki/Lanczos_algorithm http://en.m.wikipedia.org/wiki/Lanczos_algorithm