[pymvpa] making new dataset from NiftiDataset
Yaroslav Halchenko
debian at onerussian.com
Sun May 10 23:59:39 UTC 2009
there are few ways to remapping of the labels...
1. if you used literal labels originally (e.g. ['cond1', 'cond2', 'reset']
then you could as well provide labels_map={'cond1':1, 'cond2':1,
'rest':2} to the NiftiDataset constructor
2. just remember original labels somewhere
labels_all = ds.labels
and then remap them within the same dataset
remapping = {1:2, 2:1, 3:1} # if original were 1 for rest, 2,3 for conditions
ds.labels = [remapping[l] for l in ds.labels]
and then do that 'categories from rest' searchlight
or am I missing the point?
On Sun, 10 May 2009, Annalies wrote:
> Hi,
> I've run searchlight on a Nifti dataset using nu-SVM classifier and an
> odd-even splitter. Now I'd like to use searchlight to see which voxels
> do best at telling the samples from 2 categories from the rest. How do
> I do this?
> What I've tried: I've made a new dataset with label 1 for the two
> categories I'm interested in and 2 for all the other non-rest samples,
> but then I lose the mapping information searchlight needs. If I select
> samples to run the searchlight on, it will try to tell apart all the
> categories instead of just trying to distinguist categories 1 and 4
> from the others. The only way I've come up with to do this is to create
> a new numlabels file, but this is tedious and prone to error. Is there
> a better way?
> Best,
> Annalies
> _______________________________________________
> Pkg-ExpPsy-PyMVPA mailing list
> Pkg-ExpPsy-PyMVPA at lists.alioth.debian.org
> http://lists.alioth.debian.org/mailman/listinfo/pkg-exppsy-pymvpa
--
.-.
=------------------------------ /v\ ----------------------------=
Keep in touch // \\ (yoh@|www.)onerussian.com
Yaroslav Halchenko /( )\ ICQ#: 60653192
Linux User ^^-^^ [175555]
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
URL: <http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/attachments/20090510/79d53f37/attachment.pgp>
More information about the Pkg-ExpPsy-PyMVPA
mailing list