Francisco Pereira wrote:
On Tue, Apr 6, 2010 at 9:01 AM, Yaroslav Halchenko
<[email protected]> wrote:
But keep in mind -- this is GN(aive)B, so you might find results inferior
to using generic searchlight with not-so-naive multivariate learners ;)
Entirely agree with Yarik, though I'll add that across multiple
datasets we've found GNB works pretty well, close to what you get with
LDA (waiting to hear from NeuroImage on this, otherwise I could tell
you more ;). Hence it's always worth trying, especially now that it
can be done quickly.
Matthias: you can still justify the server by installing condor and
making a batch queue for all your subjects...
Thanks Francisco. I don't know condor yet (working with SGE here), but
my feelings says that it is well suited for (huge) dedicated clusters
and a little bit to "big" for my purpose... but I will definitely have a
closer look.
Matthias
_______________________________________________
Pkg-ExpPsy-PyMVPA mailing list
[email protected]
http://lists.alioth.debian.org/mailman/listinfo/pkg-exppsy-pymvpa
