File feature_vector.hpp¶

class IFeatureVector¶

#include <feature_vector.hpp>

Defines an interface for all one-dimensional vectors that store the values of training examples for a certain feature.

Public Functions

inline virtual ~IFeatureVector()¶

virtual std::unique_ptr<IResettableStatisticsSubset> createStatisticsSubset(const IWeightedStatistics &statistics, const CompleteIndexVector &outputIndices) const = 0¶

Creates and returns an IResettableStatisticsSubset based on given statistics and this feature vector.

Parameters:

statistics – A reference to an object of type IWeightedStatisticss, the subset should be based on
outputIndices – A reference to an object of type CompleteIndexVector that provides access to the indices of the outputs that should be included in the subset

Returns:

An unique pointer to an object of type IResettableStatisticsSubset that has been created

virtual std::unique_ptr<IResettableStatisticsSubset> createStatisticsSubset(const IWeightedStatistics &statistics, const PartialIndexVector &outputIndices) const = 0¶

Creates and returns an IResettableStatisticsSubset based on given statistics and this feature vector.

Parameters:

statistics – A reference to an object of type IWeightedStatisticss, the subset should be based on
outputIndices – A reference to an object of type PartialIndexVector that provides access to the indices of the outputs that should be included in the subset

Returns:

An unique pointer to an object of type IResettableStatisticsSubset that has been created

virtual void searchForRefinement(SingleRefinementComparator &comparator, const IWeightedStatistics &statistics, const IIndexVector &outputIndices, uint32 numExamplesWithNonZeroWeights, uint32 minCoverage, Refinement &refinement) const = 0¶

Conducts a search for the best refinement of an existing rule that can be created from a this feature vector.

Parameters:

comparator – A reference to an object of type SingleRefinementComparator that should be used for comparing potential refinements
statistics – A reference to an object of type IWeightedStatistics that provides access to weighted statistics about the quality of predictions for training examples, which should serve as the basis for evaluating the quality of potential refinements
outputIndices – A reference to an object of type IIndexVector that provides access to the indices of the outputs for which refinements should predict
numExamplesWithNonZeroWeights – The total number of examples with non-zero weights that may be covered by a refinement
minCoverage – The minimum number of examples that must be covered by the refinement
refinement – A reference to an object of type Refinement that should be used for storing the properties of the best refinement that is found

virtual void searchForRefinement(FixedRefinementComparator &comparator, const IWeightedStatistics &statistics, const IIndexVector &outputIndices, uint32 numExamplesWithNonZeroWeights, uint32 minCoverage, Refinement &refinement) const = 0¶

Conducts a search for the best refinement of an existing rule that can be created from a this feature vector.

Parameters:

comparator – A reference to an object of type MultiRefinementComparator that should be used for comparing potential refinements
statistics – A reference to an object of type IWeightedStatistics that provides access to weighted statistics about the quality of predictions for training examples, which should serve as the basis for evaluating the quality of potential refinements
outputIndices – A reference to an object of type IIndexVector that provides access to the indices of the outputs for which refinements should predict
numExamplesWithNonZeroWeights – The total number of examples with non-zero weights that may be covered by a refinement
minCoverage – The minimum number of examples that must be covered by the refinement
refinement – A reference to an object of type Refinement that should be used for storing the properties of the best refinement that is found

virtual void updateCoverageMaskAndStatistics(const Interval &interval, CoverageMask &coverageMask, uint32 indicatorValue, IWeightedStatistics &statistics) const = 0¶

Updates a given CoverageMask and IWeightedStatistics depending on the indices of training examples included in a specific Interval.

Parameters:

interval – A reference to an object of type Interval that includes the indices of all covered training examples
coverageMask – A reference to an object of type CoverageMask that should be updated
indicatorValue – The indicator value that should be used for updating the given CoverageMask
statistics – A reference to an object of type IWeightedStatistics that should be updated

virtual std::unique_ptr<IFeatureVector> createFilteredFeatureVector(std::unique_ptr<IFeatureVector> &existing, const Interval &interval) const = 0¶

Creates and returns a copy of this vector that does only store the feature values of training examples included in a given Interval.

Parameters:

existing – A reference to an unique pointer that stores an object of type IFeatureVector that may be reused or a null pointer, if no such object is available
interval – A reference to an object of type Interval that includes the indices of the training examples to be retained

Returns:

An unique pointer to an object of type IFeatureVector that has been created

virtual std::unique_ptr<IFeatureVector> createFilteredFeatureVector(std::unique_ptr<IFeatureVector> &existing, const CoverageMask &coverageMask) const = 0¶

Creates and returns a copy of this vector that does only store the feature values of training examples marked as covered according to a given CoverageMask.

Parameters:

existing – A reference to an unique pointer that stores an object of type IFeatureVector that may be reused or a null pointer, if no such object is available
coverageMask – A reference to an object of type CoverageMask that specifies the indices of the training examples to be retained

Returns:

An unique pointer to an object of type IFeatureVector that has been created