[pymvpa] Results from individual folds

[patrik andersson]

Hi all,

I am just getting started using PyMVPA so my question is very basic. I am
sure the answer must be available already but I just cant seem to find it.
Sorry about this, but I hope someone can help me out.
My question is how to get the classification results from the N individual
tests using the NFoldSplitter.
#Lets say I have a 10 chunks dataset 'ds' and run:
  clf = FeatureSelectionClassifier(
      LinearCSVMC(),
      SensitivityBasedFeatureSelection(
          OneWayAnova(),
      FractionTailSelector(0.05,mode='select',tail='upper')),
      descr="VA",
      enable_states = ['feature_ids'])

  terr = TransferError(clf)
  cvterr = CrossValidatedTransferError(terr,
                                     NFoldSplitter(cvtype=1),
                                     enable_states=['confusion'])

error = cvterr(ds)

#If I now do:
  print cvterr.confusion.asstring(description=True)
#it gives me one confusion matrix which I guess is the mean one over all N
folds(?)

Any hints or help is much appreciated!

Cheers,
Patrik
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/attachments/20090811/f8818307/attachment.htm>