prtOutlierRemoval Base class for prt Outlier Removal objects
 
    prtOutlierRemoval is an abstract class and cannot be instantiated.
 
    All prtOutlierRemoval objects inherit all properities and methods
    from the prtAction object. prtOutlierRemoval objects have the
    following additional properties:
 
    runMode - Specify how the outlier removal processing behaves when
    run on a data set.  runMode specifies how the action behaves during
    typical calls to RUN. Valid strings are as follows:
 
        'noAction' - When running the outlier removal action, do
        nothing.  This ensures that the outlier removal action outputs
        data sets of the same size as the input data set.
 
        'replaceWithNan' - When running the outlier removal action
        replace outlier values with nans.  This ensures that the
        outlier removal action outputs data sets of the same size as
        the input data set.
 
        'removeObservation' - When running the outlier removal action,
        remove observations where any feature value is flagged as an
        outlier.  Note: This can change the size of the data set during
        running and can result in invalid cross-validation folds.
 
        'removeFeature'  - When running the outlier removal action,
        remove features where any observation contains an outlier.
        
    runOnTrainingMode - A string specifying how the outlier removal
    method should behave when being run during the training of a
    prtAlgorithm.  See above for a more detailed description, and a
    list of valid runOnTrainingMode string values.  Default value is
    'removeObservation'.