[pymvpa] parallelizing for cross-subjects NFold cross-validation?

[Richard Dinga]

Yes, you can do it, each permutation is independent, doesn't matter if they
are computed in series or parallel. Unless you specify the same RNG seed
for your splits, there shouldn't be any problem.

BW,
Richard

On Sat, Dec 9, 2017 at 3:33 AM, Regina Lapate <lapate at gmail.com> wrote:

> Hello all:
>
> I have been trying to run an NFold cross-validation, leave-one-subject-out
> scheme, and I am writing because the process has been taking quite a while.
> Specifically, I am using SVM, n=2 labels/targets, n=130 subjects (which are
> my 'chunks'), and I initially tried 1000 permutations--but realized that
> would take weeks to run.
>
> --Do you have any suggestions of how to go about this? Would it be
> appropriate to split the job by shuffling fewer times per job, and then
> aggregate across the null distribution over those jobs at the end (or are
> the shuffles occurring in a systematic way that would render such an
> aggregation over jobs biased)?
>
> Thank you for any pointers!
>
> Regina
>
> _______________________________________________
> Pkg-ExpPsy-PyMVPA mailing list
> Pkg-ExpPsy-PyMVPA at lists.alioth.debian.org
> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/attachments/20171212/b3350765/attachment.html>