Package mvpa :: Package measures :: Module base :: Class FeaturewiseDatasetMeasure
[hide private]
[frames] | no frames]

Class FeaturewiseDatasetMeasure

source code


A per-feature-measure computed from a Dataset (base class).

Should behave like a DatasetMeasure.

Nested Classes [hide private]

Inherited from misc.state.ClassWithCollections: __metaclass__

Instance Methods [hide private]
 
__init__(self, combiner=SecondAxisSumOfAbs, **kwargs)
Initialize
source code
 
__repr__(self, prefixes=None)
String representation of DatasetMeasure
source code
 
_call(self, dataset)
Computes a per-feature-measure on a given Dataset.
source code
 
_postcall(self, dataset, result)
Adjusts per-feature-measure for computed `result`
source code
 
combiner(self)
Return combiner
source code

Inherited from DatasetMeasure: __call__, null_dist, transformer, untrain

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]
  base_sensitivities = StateVariable(enabled= False, doc= "Store...

Inherited from DatasetMeasure: __doc__, null_prob, null_t, raw_results

Inherited from misc.state.ClassWithCollections: _DEV__doc__, descr

Properties [hide private]

Inherited from object: __class__

Method Details [hide private]

__init__(self, combiner=SecondAxisSumOfAbs, **kwargs)
(Constructor)

source code 
Initialize
Parameters:
  • combiner (Functor) - The combiner is only applied if the computed featurewise dataset measure is more than one-dimensional. This is different from a transformer, which is always applied. By default, the sum of absolute values along the second axis is computed.
Overrides: object.__init__

__repr__(self, prefixes=None)
(Representation operator)

source code 

String representation of DatasetMeasure

Includes only arguments which differ from default ones

Parameters:
  • fullname - Either to include full name of the module
  • prefixes - What other prefixes to prepend to list of arguments
Overrides: object.__repr__
(inherited documentation)

_call(self, dataset)

source code 

Computes a per-feature-measure on a given Dataset.

Behaves like a DatasetMeasure, but computes and returns a 1d ndarray with one value per feature.

Overrides: DatasetMeasure._call

_postcall(self, dataset, result)

source code 
Adjusts per-feature-measure for computed `result`


TODO: overlaps in what it does heavily with
 CombinedSensitivityAnalyzer, thus this one might make use of
 CombinedSensitivityAnalyzer yoh thinks, and here
 base_sensitivities doesn't sound appropriate.
 MH: There is indeed some overlap, but also significant differences.
     This one operates on a single sensana and combines over second
     axis, CombinedFeaturewiseDatasetMeasure uses first axis.
     Additionally, 'Sensitivity' base class is
     FeaturewiseDatasetMeasures which would have to be changed to
     CombinedFeaturewiseDatasetMeasure to deal with stuff like
     SMLRWeights that return multiple sensitivity values by default.
     Not sure if unification of both (and/or removal of functionality
     here does not lead to an overall more complicated situation,
     without any real gain -- after all this one works ;-)

Overrides: DatasetMeasure._postcall

combiner(self)

source code 
Return combiner
Decorators:
  • @property

Class Variable Details [hide private]

base_sensitivities

Value:
StateVariable(enabled= False, doc= "Stores basic sensitivities if the \
sensitivity "+ "relies on combining multiple ones")