def __init__(self,
data_measure,
transfer_error,
bestdetector=BestDetector(),
stopping_criterion=NBackHistoryStopCrit(BestDetector()),
feature_selector=FixedNElementTailSelector(1,
tail='upper',
mode='select'),
**kwargs
):
"""Initialize incremental feature search
Parameters
----------
data_measure : DatasetMeasure
Computed for each candidate feature selection. The measure has
to compute a scalar value.
transfer_error : TransferError
Compute against a test dataset for each incremental feature
set.
bestdetector : Functor
Given a list of error values it has to return a boolean that
signals whether the latest error value is the total minimum.
stopping_criterion : Functor
Given a list of error values it has to return whether the
criterion is fulfilled.
"""
# bases init first
FeatureSelection.__init__(self, **kwargs)
self.__data_measure = data_measure
self.__transfer_error = transfer_error
self.__feature_selector = feature_selector
self.__bestdetector = bestdetector
self.__stopping_criterion = stopping_criterion
def __init__(self,
data_measure,
transfer_error,
bestdetector=BestDetector(),
stopping_criterion=NBackHistoryStopCrit(BestDetector()),
feature_selector=FixedNElementTailSelector(1,
tail='upper',
mode='select'),
**kwargs):
"""Initialize incremental feature search
Parameters
----------
data_measure : DatasetMeasure
Computed for each candidate feature selection. The measure has
to compute a scalar value.
transfer_error : TransferError
Compute against a test dataset for each incremental feature
set.
bestdetector : Functor
Given a list of error values it has to return a boolean that
signals whether the latest error value is the total minimum.
stopping_criterion : Functor
Given a list of error values it has to return whether the
criterion is fulfilled.
"""
# bases init first
FeatureSelection.__init__(self, **kwargs)
self.__data_measure = data_measure
self.__transfer_error = transfer_error
self.__feature_selector = feature_selector
self.__bestdetector = bestdetector
self.__stopping_criterion = stopping_criterion
def __init__(self,val=0,n_select=0.05):
FeatureSelection.__init__(self)
self.__ttfm = TTestFeaturewiseMeasure(val)
self._combinedSA = CombinedFeaturewiseDatasetMeasure(analyzers=[self.__ttfm,OneWayAnovaProbs()], combiner = np.prod)
if n_select<1 and n_select>0:
self._selector=FractionTailSelector(n_select, mode='select', tail='lower')
elif n_select>1:
self._selector=FixedNElementTailSelector(int(n_select), mode='select', tail='lower')
else:
raise ValueError("n_select must be > 0, is %f"%n_select)
self.__fs = SensitivityBasedFeatureSelection(
self._combinedSA,
self._selector,
enable_states = ["sensitivity", "selected_ids"],
)
def __init__(self,
test_type = "ks",
num_features = 10,
**kwargs
):
"""Initialize feature selection
:Parameters:
test_type : 'anova', 'ttest', 'ks', 'svm'
type of statistical test to perform
num_features :
Given a sensitivity map it has to return the ids of those
features that should be kept.
"""
# base init first
FeatureSelection.__init__(self, **kwargs)
self.__test_type = test_type
"""Test-type, One of 'anova', 'ttest', 'ks', 'svm'"""
self.__num_features = num_features
"""Number of features to include"""
def __init__(self,
sensitivity_analyzer,
transfer_error,
feature_selector=FractionTailSelector(0.05),
bestdetector=BestDetector(),
stopping_criterion=NBackHistoryStopCrit(BestDetector()),
train_clf=None,
update_sensitivity=True,
**kargs
):
# XXX Allow for multiple stopping criterions, e.g. error not decreasing
# anymore OR number of features less than threshold
"""Initialize recursive feature elimination
:Parameters:
sensitivity_analyzer : FeaturewiseDatasetMeasure object
transfer_error : TransferError object
used to compute the transfer error of a classifier based on a
certain feature set on the test dataset.
NOTE: If sensitivity analyzer is based on the same
classifier as transfer_error is using, make sure you
initialize transfer_error with train=False, otherwise
it would train classifier twice without any necessity.
feature_selector : Functor
Given a sensitivity map it has to return the ids of those
features that should be kept.
bestdetector : Functor
Given a list of error values it has to return a boolean that
signals whether the latest error value is the total minimum.
stopping_criterion : Functor
Given a list of error values it has to return whether the
criterion is fulfilled.
train_clf : bool
Flag whether the classifier in `transfer_error` should be
trained before computing the error. In general this is
required, but if the `sensitivity_analyzer` and
`transfer_error` share and make use of the same classifier it
can be switched off to save CPU cycles. Default `None` checks
if sensitivity_analyzer is based on a classifier and doesn't train
if so.
update_sensitivity : bool
If False the sensitivity map is only computed once and reused
for each iteration. Otherwise the senstitivities are
recomputed at each selection step.
"""
# base init first
FeatureSelection.__init__(self, **kargs)
self.__sensitivity_analyzer = sensitivity_analyzer
"""Sensitivity analyzer used to call at each step."""
self.__transfer_error = transfer_error
"""Compute transfer error for each feature set."""
self.__feature_selector = feature_selector
"""Functor which takes care about removing some features."""
self.__stopping_criterion = stopping_criterion
self.__bestdetector = bestdetector
if train_clf is None:
self.__train_clf = isinstance(sensitivity_analyzer,
Sensitivity)
else:
self.__train_clf = train_clf
"""Flag whether training classifier is required."""
self.__update_sensitivity = update_sensitivity
"""Flag whether sensitivity map is recomputed for each step."""
# force clf training when sensitivities are not updated as otherwise
# shared classifiers are not retrained
if not self.__update_sensitivity \
and isinstance(self.__transfer_error, ClassifierError) \
and not self.__train_clf:
if __debug__:
debug("RFEC", "Forcing training of classifier since " +
"sensitivities aren't updated at each step")
self.__train_clf = True
def __init__(self,
sensitivity_analyzer,
transfer_error,
feature_selector=FractionTailSelector(0.05),
bestdetector=BestDetector(),
stopping_criterion=NBackHistoryStopCrit(BestDetector()),
train_clf=None,
update_sensitivity=True,
**kargs
):
# XXX Allow for multiple stopping criterions, e.g. error not decreasing
# anymore OR number of features less than threshold
"""Initialize recursive feature elimination
Parameters
----------
sensitivity_analyzer : FeaturewiseDatasetMeasure object
transfer_error : TransferError object
used to compute the transfer error of a classifier based on a
certain feature set on the test dataset.
NOTE: If sensitivity analyzer is based on the same
classifier as transfer_error is using, make sure you
initialize transfer_error with train=False, otherwise
it would train classifier twice without any necessity.
feature_selector : Functor
Given a sensitivity map it has to return the ids of those
features that should be kept.
bestdetector : Functor
Given a list of error values it has to return a boolean that
signals whether the latest error value is the total minimum.
stopping_criterion : Functor
Given a list of error values it has to return whether the
criterion is fulfilled.
train_clf : bool
Flag whether the classifier in `transfer_error` should be
trained before computing the error. In general this is
required, but if the `sensitivity_analyzer` and
`transfer_error` share and make use of the same classifier it
can be switched off to save CPU cycles. Default `None` checks
if sensitivity_analyzer is based on a classifier and doesn't train
if so.
update_sensitivity : bool
If False the sensitivity map is only computed once and reused
for each iteration. Otherwise the senstitivities are
recomputed at each selection step.
"""
# base init first
FeatureSelection.__init__(self, **kargs)
self.__sensitivity_analyzer = sensitivity_analyzer
"""Sensitivity analyzer used to call at each step."""
self.__transfer_error = transfer_error
"""Compute transfer error for each feature set."""
self.__feature_selector = feature_selector
"""Functor which takes care about removing some features."""
self.__stopping_criterion = stopping_criterion
self.__bestdetector = bestdetector
if train_clf is None:
self.__train_clf = isinstance(sensitivity_analyzer,
Sensitivity)
else:
self.__train_clf = train_clf
"""Flag whether training classifier is required."""
self.__update_sensitivity = update_sensitivity
"""Flag whether sensitivity map is recomputed for each step."""
# force clf training when sensitivities are not updated as otherwise
# shared classifiers are not retrained
if not self.__update_sensitivity \
and isinstance(self.__transfer_error, ClassifierError) \
and not self.__train_clf:
if __debug__:
debug("RFEC", "Forcing training of classifier since " +
"sensitivities aren't updated at each step")
self.__train_clf = True
