Package mvpa :: Package algorithms :: Module cvtranserror :: Class CrossValidatedTransferError
[hide private]
[frames] | no frames]

Class CrossValidatedTransferError

source code


Classifier cross-validation.

This class provides a simple interface to cross-validate a classifier on datasets generated by a splitter from a single source dataset.

Arbitrary performance/error values can be computed by specifying an error function (used to compute an error value for each cross-validation fold) and a combiner function that aggregates all computed error values across cross-validation folds.

Nested Classes [hide private]

Inherited from misc.state.ClassWithCollections: __metaclass__

Instance Methods [hide private]
 
__init__(self, transerror, splitter=None, combiner='mean', expose_testdataset=False, harvest_attribs=None, copy_attribs='copy', **kwargs)
:Parameters: transerror: TransferError instance Provides the classifier used for cross-validation.
source code
 
_call(self, dataset)
Perform cross-validation on a dataset.
source code

Inherited from measures.base.DatasetMeasure: __call__, __repr__, null_dist, transformer, untrain

Inherited from measures.base.DatasetMeasure (private): _postcall

Inherited from misc.state.Harvestable (private): _harvest, _setAttribs

Inherited from misc.state.ClassWithCollections: __getattribute__, __new__, __setattr__, __str__, reset

Inherited from object: __delattr__, __format__, __hash__, __reduce__, __reduce_ex__, __sizeof__, __subclasshook__

Class Variables [hide private]
  results = StateVariable(enabled= False, doc= """Store individu...
  splits = StateVariable(enabled= False, doc= """Store the actua...
  transerrors = StateVariable(enabled= False, doc= """Store copi...
  confusion = StateVariable(enabled= False, doc= """Store total ...
  training_confusion = StateVariable(enabled= False, doc= """Sto...
  samples_error = StateVariable(enabled= False, doc= "Per sample...
  splitter = property(fget= lambda self: self.__splitter, doc= "...
  transerror = property(fget= lambda self: self.__transerror, do...
  combiner = property(fget= lambda self: self.__combiner, doc= "...

Inherited from measures.base.DatasetMeasure: __doc__, null_prob, null_t, raw_results

Inherited from misc.state.Harvestable: harvest_attribs, harvested

Inherited from misc.state.Harvestable (private): _KNOWN_COPY_METHODS

Inherited from misc.state.ClassWithCollections: _DEV__doc__, descr

Properties [hide private]

Inherited from object: __class__

Method Details [hide private]

__init__(self, transerror, splitter=None, combiner='mean', expose_testdataset=False, harvest_attribs=None, copy_attribs='copy', **kwargs)
(Constructor)

source code 

:Parameters:
  transerror: TransferError instance
    Provides the classifier used for cross-validation.
  splitter: Splitter | None
    Used to split the dataset for cross-validation folds. By
    convention the first dataset in the tuple returned by the
    splitter is used to train the provided classifier. If the
    first element is 'None' no training is performed. The second
    dataset is used to generate predictions with the (trained)
    classifier. If `None` (default) an instance of
    :class:`~mvpa.datasets.splitters.NoneSplitter` is used.
  combiner: Functor | 'mean'
    Used to aggregate the error values of all cross-validation
    folds. If 'mean' (default) the grand mean of the transfer
    errors is computed.
  expose_testdataset: bool
   In the proper pipeline, classifier must not know anything
   about testing data, but in some cases it might lead only
   to marginal harm, thus migth wanted to be enabled (provide
   testdataset for RFE to determine stopping point).
  harvest_attribs: list of basestr
    What attributes of call to store and return within
    harvested state variable
  copy_attribs: None | basestr
    Force copying values of attributes on harvesting
  **kwargs:
    All additional arguments are passed to the
    :class:`~mvpa.measures.base.DatasetMeasure` base class.

Parameters:
  • transformer, Functor - This functor is called in __call__() to perform a final processing step on the to be returned dataset measure. If None, nothing is called
  • null_dist, instance, of, distribution, estimator - The estimated distribution is used to assign a probability for a certain value of the computed measure.
Overrides: object.__init__

_call(self, dataset)

source code 

Perform cross-validation on a dataset.

'dataset' is passed to the splitter instance and serves as the source dataset to generate split for the single cross-validation folds.

Overrides: measures.base.DatasetMeasure._call

Class Variable Details [hide private]

results

Value:
StateVariable(enabled= False, doc= """Store individual results in the \
state""")

splits

Value:
StateVariable(enabled= False, doc= """Store the actual splits of the d\
ata. Can be memory expensive""")

transerrors

Value:
StateVariable(enabled= False, doc= """Store copies of transerrors at e\
ach step. If enabled -
       operates on clones of transerror, but for the last split origin\
al
       transerror is used""")

confusion

Value:
StateVariable(enabled= False, doc= """Store total confusion matrix (if\
 available)""")

training_confusion

Value:
StateVariable(enabled= False, doc= """Store total training confusion m\
atrix (if available)""")

samples_error

Value:
StateVariable(enabled= False, doc= "Per sample errors.")

splitter

Value:
property(fget= lambda self: self.__splitter, doc= "Access to the Split\
ter instance.")

transerror

Value:
property(fget= lambda self: self.__transerror, doc= "Access to the Tra\
nsferError instance.")

combiner

Value:
property(fget= lambda self: self.__combiner, doc= "Access to the confi\
gured combiner.")