This content refers to the previous stable release of PyMVPA. Please visit www.pymvpa.org for the most recent version of PyMVPA and its documentation.

featsel.ifs

Module: featsel.ifs

Inheritance diagram for mvpa.featsel.ifs:

Incremental feature search (IFS).

Very similar to Recursive feature elimination (RFE), but instead of begining with all features and stripping some sequentially, start with an empty feature set and include important features successively.

IFS

class mvpa.featsel.ifs.IFS(data_measure, transfer_error, bestdetector=<mvpa.featsel.helpers.BestDetector object at 0x99868d0>, stopping_criterion=<mvpa.featsel.helpers.NBackHistoryStopCrit object at 0x9986e90>, feature_selector=FixedNElementTailSelector() number=1.000000, **kwargs)

Bases: mvpa.featsel.base.FeatureSelection

Incremental feature search.

A scalar DatasetMeasure is computed multiple times on variations of a certain dataset. These measures are in turn used to incrementally select important features. Starting with an empty feature set the dataset measure is first computed for each single feature. A number of features is selected based on the resulting data measure map (using an ElementSelector).

Next the dataset measure is computed again using each feature in addition to the already selected feature set. Again the ElementSelector is used to select more features.

For each feature selection the transfer error on some testdatset is computed. This procedure is repeated until a given StoppingCriterion is reached.

Note

Available state variables:

• errors+: State variable
• selected_ids: State variable

(States enabled by default are listed with +)

See also

Please refer to the documentation of the base class for more information:

FeatureSelection

Initialize incremental feature search

Parameters:

• data_measure (DatasetMeasure) – Computed for each candidate feature selection.
• 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.
• enable_states (None or list of basestring) – Names of the state variables which should be enabled additionally to default ones
• disable_states (None or list of basestring) – Names of the state variables which should be disabled