[pymvpa] SVM + RFE

[Arman Eshaghi]

Dear all,

I'm struggling with recursive feature selection, a meta classifier (clf),
and the final cross validation. Below is what I have done and what I get as
error. I would very much appreciate if you could help me here.

*#from rfe manual*

*rfe* = RFE(rfesvm_split.get_sensitivity_analyzer(postproc = ChainMapper([
FxMapper('features', l2_normed), FxMapper('samples', np.mean),
FxMapper('samples', np.abs)])), ConfusionBasedError(rfesvm_split,
confusion_state='stats'),
Repeater(2), fselector=FractionTailSelector(0.50, mode = 'select', tail=
'upper'), stopping_criterion=NBackHistoryStopCrit(BestDetector(), 10),
train_pmeasure=False, update_sensitivity=True)

*#Meta-classifier with SVM as the final classifier and rfe as feature
selector*

*clf*=FeatureSelectionClassifier(LinearCSVMC(), rfe, descr='SVM+RFE')

*#cross-validation*

*cvte*=CrossValidation(clf, HalfPartitioner(), enable_ca=['stats'])

*#running the analysis*

*results*=cvte(MyData)

*ERRORS:*

In [37]: results=cvte(gm_lt)
---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-37-a9da6f9cc192> in <module>()
*----> 1 results=cvte(gm_lt)*

/usr/lib64/python2.6/site-packages/mvpa2/base/learner.pyc in __call__(self,
ds)
    257                                    "used and auto training is
disabled."
    258                                    % str(self))
*--> 259         return super(Learner, self).__call__(ds)*
    260
    261

/usr/lib64/python2.6/site-packages/mvpa2/base/node.pyc in __call__(self, ds)
    109
    110         self._precall(ds)
*--> 111         result = self._call(ds)*
    112         result = self._postcall(ds, result)
    113

/usr/lib64/python2.6/site-packages/mvpa2/measures/base.pyc in _call(self,
ds)
    495         # always untrain to wipe out previous stats
    496         self.untrain()
*--> 497         return super(CrossValidation, self)._call(ds)*
    498
    499

/usr/lib64/python2.6/site-packages/mvpa2/measures/base.pyc in _call(self,
ds)
    324                 ca.datasets.append(sds)
    325             # run the beast
*--> 326             result = node(sds)*
    327             # callback
    328             if not self._callback is None:

/usr/lib64/python2.6/site-packages/mvpa2/base/learner.pyc in __call__(self,
ds)
    257                                    "used and auto training is
disabled."
    258                                    % str(self))
*--> 259         return super(Learner, self).__call__(ds)*
    260
    261

/usr/lib64/python2.6/site-packages/mvpa2/base/node.pyc in __call__(self, ds)
    109
    110         self._precall(ds)
*--> 111         result = self._call(ds)*
    112         result = self._postcall(ds, result)
    113

/usr/lib64/python2.6/site-packages/mvpa2/measures/base.pyc in _call(self,
ds)
    589                     for i in
dstrain.get_attr(splitter.get_space())[0].unique])
    590         # ask splitter for first part
*--> 591         measure.train(dstrain)*
    592         # cleanup to free memory
    593         del dstrain

/usr/lib64/python2.6/site-packages/mvpa2/base/learner.pyc in train(self, ds)
    130             # things might have happened during pretraining
    131             if ds.nfeatures > 0:
*--> 132                 result = self._train(ds)*
    133             else:
    134                 warning("Trying to train on dataset with no
features present")

/usr/lib64/python2.6/site-packages/mvpa2/clfs/meta.pyc in _train(self,
dataset)
   1346         # XXX: should training be done using whole dataset or just
samples
   1347         # YYY: in some cases labels might be needed, thus better
full dataset
*-> 1348         self.__mapper.train(dataset)*
   1349
   1350         # for train() we have to provide dataset -- not just
samples to train!

/usr/lib64/python2.6/site-packages/mvpa2/base/learner.pyc in train(self, ds)
    130             # things might have happened during pretraining
    131             if ds.nfeatures > 0:
*--> 132                 result = self._train(ds)*
    133             else:
    134                 warning("Trying to train on dataset with no
features present")

/usr/lib64/python2.6/site-packages/mvpa2/featsel/rfe.pyc in _train(self, ds)
    246             # Compute sensitivity map
    247             if self.__update_sensitivity or sensitivity == None:
*--> 248                 sensitivity = self._fmeasure(wdataset)*
    249                 if len(sensitivity) > 1:
    250                     raise ValueError(

/usr/lib64/python2.6/site-packages/mvpa2/base/learner.pyc in __call__(self,
ds)
    251                     debug('LRN', "Auto-training %s on %s",
    252                           (self, ds))
*--> 253                 self.train(ds)*
    254             else:
    255                 # we always have to have trained before using a
learner

/usr/lib64/python2.6/site-packages/mvpa2/base/learner.pyc in train(self, ds)
    130             # things might have happened during pretraining
    131             if ds.nfeatures > 0:
*--> 132                 result = self._train(ds)*
    133             else:
    134                 warning("Trying to train on dataset with no
features present")

/usr/lib64/python2.6/site-packages/mvpa2/measures/base.pyc in _train(self,
dataset)
    805                     True:  "although it was trained previously"}
    806                    [clf.trained]))
*--> 807         return clf.train(dataset)*
    808
    809

/usr/lib64/python2.6/site-packages/mvpa2/base/learner.pyc in train(self, ds)
    130             # things might have happened during pretraining
    131             if ds.nfeatures > 0:
*--> 132                 result = self._train(ds)*
    133             else:
    134                 warning("Trying to train on dataset with no
features present")

/usr/lib64/python2.6/site-packages/mvpa2/clfs/meta.pyc in _train(self,
dataset)
   1268
   1269             if ca.is_enabled("stats"):
*-> 1270                 predictions = clf.predict(split[1])*
   1271
self.ca.stats.add(split[1].sa[targets_sa_name].value,
   1272                                           predictions,

*IndexError: list index out of range*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/attachments/20140127/9c09e109/attachment.html>