auswahl.FeatureDescriptor

class auswahl.FeatureDescriptor(key: Union[int, Tuple[int, int], FeatureDescriptor], resolve_intervals: bool = False)

The class FeatureDescriptor abstracts the configuration of features the selection methods are to retrieve from the spectral data. The FeatureDescriptor wraps either a number of arbitrary features to be selected or a specific number of intervals of features of a fix length.

Parameters

key: int, Tuple[int, int], FeatureDescriptor

Feature configuration to be abstracted by the object. A single integer is interpreted as a number of arbitrarily selectable features. A tuple is a (#intervals, width of intervals) configuration of features to be selected. If a FeatureDescriptor is passed, it is copied. All passed integers are required to be non-negative.

resolve_intervals: bool, default=False

Flag indicating whether interval feature configurations are to be resolved to a single integer of arbitrary features to be selected.

Attributes

org_key: int, Tuple[int, int]

Originally passed feature configuration

key: int, Tuple[int, int]

Resolved key. Equal to org_key, if org_key is not a tuple or if argument resolve_tuples is False

resolve_intervals: bool

Passed argument resolve_intervals

__init__(key: Union[int, Tuple[int, int], FeatureDescriptor], resolve_intervals: bool = False)

property comparator

Provides a feature configuration representation allowing comparison of FeatureDescriptors.

get_configuration_for(selector: SpectralSelector)

Translate and return the feature configuration for a given SpectralSelector.

Parameters

selector: SpectralSelector

SpectralSelector instance

property string_rep

Provides a printing representation for the FeatureDescriptor printing interval configurations as number of intervals and interval width separated via a forward slash.

Returns

string representation: str