[Pkg-exppsy-pymvpa] model selection and other questions

[Yaroslav Halchenko]

On Mon, 07 Apr 2008, Emanuele Olivetti wrote:

> > Thanks for sharing the experience -- we should give them a spin too but
> > lots of changes should happen first to make that viable  -- for
> > instance, now considerable amount of 'training' of SVMs is taken by
> > transforming data into 'library' space.
> I don't get the point, but I'll catch it eventually, since I'm playing
> with pyMVPA now.
just look for convert2SVMNode in case of libsvm.
I believe that scikits.learn had it slightly more effective, but I might
be wrong

In case of shogun -- it is not only that: shogun precomputes kernels'
matrix, thus if datasamples didn't change, it is not needed to recompute
it if only parameters of SVM  (not of the kernel) has changed (ie C for
soft-margin C-SVM). 

> Bye,

> Emanuele



-- 
Yaroslav Halchenko
Research Assistant, Psychology Department, Rutgers-Newark
Student  Ph.D. @ CS Dept. NJIT
Office: (973) 353-5440x263 | FWD: 82823 | Fax: (973) 353-1171
        101 Warren Str, Smith Hall, Rm 4-105, Newark NJ 07102
WWW:     http://www.linkedin.com/in/yarik