On Tue, Mar 27, 2012 at 4:41 PM, Dan Brickley <[email protected]> wrote: > Maybe this is as good an excuse as any to get my hands dirty with > SSVD... Is there any strong reason to prefer Lanczos for this sort of > thing? Depends on app? (mine won't care massively about super > precision; speed would generally be more of a concern)
Reasons to prefer SSVD are mainly speed and higher eigenvalue count. Also, trade-offs speed/precision is adjustable with -q and -p parameters. With q=1 i don't notice significant difference in eigenvalues (significant as in by eyeballing the plots of such values) and it looks like SSVD is also preferrable even due to precision beyond 40-th eigenvalue on a wikipedia dataset. > > > cheers, > > > Dan
