[pymvpa] Searchligh SVM on fMRI Beta Values in PyMVPA (Yaroslav O Halchenko)

Nick Oosterhof n.n.oosterhof at googlemail.com
Sun May 26 17:01:24 BST 2019

On Sun, 26 May 2019 at 11:17, Alireza Karami <alireza.kr at gmail.com> wrote:

> I have done it! I loaded beta values and did searchlight SVM based on the
> example <http://www.pymvpa.org/examples/searchlight.html> on PyMVPA site.
> The following is my code:
>> [...]
> but when I run my code, I get the following message:
>> [SLC] DBG:                        +0:04:48 _______[0%]_______
>> -347+16:07:18  ROI 2 (2/208302), 57 features
> I think it takes time more than usual and I think something is wrong in my
> code, but I do not know what is wrong with it.

Did you already run a similar analysis in CoSMoMVPA before?

I noticed that you have 320 samples, that's quite a lot if these are beta
estimates (I assume from running a GLM?).

Some reasons why the analysis takes long:

- many chunks, thus many folds. How many chunks do you have?
- the SVM classifier may take long to converge.

If you run the analysis with a HalfPartitioner and another classifier (e.g.
nearest neighbour) does it still take a long time?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/pkg-exppsy-pymvpa/attachments/20190526/75b50553/attachment.html>

More information about the Pkg-ExpPsy-PyMVPA mailing list