[pymvpa] New Elastic Net classifier

[patrik andersson]

Hi,
I am trying to get the ENET working but get some error. I guess its a
rpy problem (?), but thought I'll take a chance and reply to this
thread.
Code snippet:

clf = FeatureSelectionClassifier(
        ENET(lm=1.0,trace=False,normalize=False),
        SensitivityBasedFeatureSelection(
            OneWayAnova(),
            FractionTailSelector(0.05,mode='select',tail='upper')),
        descr="apa",
        enable_states = ['feature_ids'])

terr = TransferError(clf)
cvterr = CrossValidatedTransferError(terr,
                                     NFoldSplitter(cvtype=5),
                                     enable_states=['confusion','harvested'])

error = cvterr(ds)

gives:

Traceback (most recent call last):
  File "C:\Project\PatternAnalysis\PyMVPA\EgnaScripts\test2.py", line
92, in <module>
    error = cvterr(ds)
  File "C:\Python25\Lib\site-packages\mvpa\measures\base.py", line
105, in __call__
    result = self._call(dataset)
  File "C:\Python25\Lib\site-packages\mvpa\algorithms\cvtranserror.py",
line 173, in _call
    result = transerror(split[1], split[0])
  File "C:\Python25\Lib\site-packages\mvpa\clfs\transerror.py", line
1283, in __call__
    self._precall(testdataset, trainingdataset)
  File "C:\Python25\Lib\site-packages\mvpa\clfs\transerror.py", line
1239, in _precall
    self.__clf.train(trainingdataset)
  File "C:\Python25\Lib\site-packages\mvpa\clfs\base.py", line 354, in train
    result = self._train(dataset)
  File "C:\Python25\Lib\site-packages\mvpa\clfs\meta.py", line 1083, in _train
    self.__maskclf.clf.train(wdataset)
  File "C:\Python25\Lib\site-packages\mvpa\clfs\base.py", line 354, in train
    result = self._train(dataset)
  File "C:\Python25\Lib\site-packages\mvpa\clfs\enet.py", line 133, in _train
    trace=self.__trace)
TypeError: object of type 'numpy.float32' has no len()


Any help is very appreciated! (Hopefully some day I will be on the
other side of the Q&A ;) )

PS. I have Vista and R 2.6.2

Kind regards,
Patrik