Classification refinement builds up the combined grid of all classes, then iterates through it always keeping track of the neighbors to detect class changes between neighbors. More...

#include <ClassificationRefinementFunctor.hpp>

Inheritance diagram for sgpp::datadriven::ClassificationRefinementFunctor:

Public Member Functions
std::vector< std::vector< size_t > >	adaptAllGrids ()
	Refine and Coarsen all grids of the model.

	ClassificationRefinementFunctor (std::vector< base::Grid * > grids, std::vector< base::DataVector * > alphas, std::vector< double > priors, size_t refinements_num=1, size_t coarsenings_num=1, bool level_penalize=true, sgpp::base::AdaptivityThresholdType thresholdType=sgpp::base::AdaptivityThresholdType::Absolute, double refinementThreshold=0.0, double coarseningThreshold=1.0, bool coarsenInitialPoints=true, size_t minimumCoarseningIndex=0)
	Constructor.

size_t	getNumGrids () override
	Returns the number of grids the functor can / does refine.

size_t	getRefinementsNum () const override
	Returns the maximal number of points that should be refined.

double	getRefinementThreshold () const override
	Returns the threshold for refinement.

double	operator() (base::GridStorage &storage, size_t seq) const override
	This should be returning a refinement value for every grid point.

void	preComputeEvaluations () override
	Used if expensive computations (eg.

void	setGridIndex (size_t grid_index) override
	Sets the index (into the vector of grids) of the grid to be refined.

double	start () const override
	Returns the lower bound of refinement criterion (e.g., alpha or error) (lower bound).

	~ClassificationRefinementFunctor () override

Public Member Functions inherited from sgpp::datadriven::MultiGridRefinementFunctor
virtual	~MultiGridRefinementFunctor ()

Public Member Functions inherited from sgpp::base::RefinementFunctor
virtual double	getTotalRefinementValue (GridStorage &storage) const
	Returns the total sum of local (error) indicators used for refinement.

	RefinementFunctor ()
	Constructor.

virtual	~RefinementFunctor ()
	Destructor.

Protected Member Functions
void	collectNeighbors (base::HashGridPoint leaf, base::HashGridPoint neighbor, size_t dim, bool isLeft)
	Stores the relation of two found neighbors.

std::vector< double >	getEvalVector (size_t ind, size_t seq) const
	Gets the evaluations of all grids at the coords of seq.

void	insertCoarseningCandidate (size_t y, std::vector< std::multimap< double, std::tuple< size_t, size_t, bool > > > &classMapsCoarsening, size_t leafSeqNumber, double score, std::tuple< size_t, size_t, bool > candidate)
	Checks if a refinement candidate should be inserted into the map of candidates.

int	sgn (double d) const

void	stepDown (size_t d, size_t minDim, base::HashGridPoint &gp, std::vector< std::pair< base::HashGridPoint, base::HashGridPoint > > &neighbors)
	Iterate through the sparse grid tree.

Protected Attributes
std::vector< base::DataVector * >	alphas

size_t	coarsenings_num
	Maximum number of coarsened points per iteration.

double	coarseningThreshold

bool	coarsenInitialPoints
	Whether or not to coarsen initial grid points.

std::vector< base::Grid * >	grids

bool	level_penalize

size_t	minimumCoarseningIndex

std::vector< std::tuple< size_t, size_t, size_t, bool, base::GridPoint::level_type > >	neighborRels

std::vector< std::map< std::string, double > >	pre_comp_evals
	Stores grid evaluations at all grids (vector) at the union of grid point coordinates over all grids (hashed by string representation in the map)

std::vector< double >	priors

size_t	refinements_num

double	refinementThreshold

sgpp::base::AdaptivityThresholdType	thresholdType

base::GridStorage	total_grid
	Maximum index of the initial grid points.

Additional Inherited Members
Public Types inherited from sgpp::base::RefinementFunctor
typedef double	value_type

Detailed Description

Classification refinement builds up the combined grid of all classes, then iterates through it always keeping track of the neighbors to detect class changes between neighbors.

Only the children in the relevant dimension and direction will be added to the grid.

Constructor & Destructor Documentation

◆ ClassificationRefinementFunctor()

sgpp::datadriven::ClassificationRefinementFunctor::ClassificationRefinementFunctor	(	std::vector< base::Grid * >	grids,
		std::vector< base::DataVector * >	alphas,
		std::vector< double >	priors,
		size_t	refinements_num = `1`,
		size_t	coarsenings_num = `1`,
		bool	level_penalize = `true`,
		sgpp::base::AdaptivityThresholdType	thresholdType = `sgpp::base::AdaptivityThresholdType::Absolute`,
		double	refinementThreshold = `0.0`,
		double	coarseningThreshold = `1.0`,
		bool	coarsenInitialPoints = `true`,
		size_t	minimumCoarseningIndex = `0`
	)

Constructor.

Parameters

grids	Vector of grids. current_grid_index specifies the grid to be refined
alphas	Vector of surpluses related to the grids
priors	Vector of priors related to the classificator
refinements_num	Maximum number of refinements done
coarsenings_num	Maximum number of coarsenings done
level_penalize	If a level penalizing is multiplied to the score (2^{\|l\|_1})
thresholdType	type of the threshold: absolute score value or relative score value
refinementThreshold	Threshold for refinement scores
coarseningThreshold	Threshold for coarsening scores
coarsenInitialPoints	Whether or not to coarsen initial grid points
minimumCoarseningIndex	Mimimum index of grid points that can be coarsened, used when coarsenInitialPoints = false.

◆ ~ClassificationRefinementFunctor()

sgpp::datadriven::ClassificationRefinementFunctor::~ClassificationRefinementFunctor ( )

inlineoverride

Member Function Documentation

◆ adaptAllGrids()

std::vector< std::vector< size_t > > sgpp::datadriven::ClassificationRefinementFunctor::adaptAllGrids ( )

Refine and Coarsen all grids of the model.

Returns: vector that for each class contains a list of indices of deleted grid points

References alphas, coarsenings_num, coarseningThreshold, coarsenInitialPoints, sgpp::op_factory::createOperationMultipleEval(), sgpp::base::OperationMultipleEval::eval(), sgpp::base::DataMatrix::get(), sgpp::base::HashGridPoint::getLeftChild(), sgpp::base::HashGridStorage::getPoint(), sgpp::base::HashGridPoint::getRightChild(), sgpp::base::HashGridPoint::getRoot(), sgpp::base::HashGridStorage::getSize(), sgpp::base::HashGridPoint::getStandardCoordinates(), grids, sgpp::base::HashGridStorage::insert(), insertCoarseningCandidate(), sgpp::base::HashGridStorage::isContaining(), minimumCoarseningIndex, sgpp::base::DataVector::mult(), neighborRels, priors, sgpp::base::HashGridStorage::recalcLeafProperty(), refinements_num, refinementThreshold, sgpp::base::HashGridPoint::rehash(), sgpp::base::Relative, sgpp::base::HashGridPoint::set(), sgpp::base::DataMatrix::setColumn(), sgpp::base::DataMatrix::setRow(), stepDown(), thresholdType, and total_grid.

Referenced by sgpp::datadriven::ModelFittingClassification::adapt().

◆ collectNeighbors()

void sgpp::datadriven::ClassificationRefinementFunctor::collectNeighbors	(	base::HashGridPoint	leaf,
		base::HashGridPoint	neighbor,
		size_t	dim,
		bool	isLeft
	)

protected

Stores the relation of two found neighbors.

References sgpp::base::HashGridPoint::getLevelSum(), sgpp::base::HashGridStorage::getSequenceNumber(), sgpp::base::HashGridStorage::insert(), sgpp::base::HashGridStorage::isContaining(), neighborRels, and total_grid.

Referenced by stepDown().

◆ getEvalVector()

std::vector< double > sgpp::datadriven::ClassificationRefinementFunctor::getEvalVector	(	size_t	ind,
		size_t	seq
	)		const

protected

Gets the evaluations of all grids at the coords of seq.

◆ getNumGrids()

size_t sgpp::datadriven::ClassificationRefinementFunctor::getNumGrids ( )

overridevirtual

Returns the number of grids the functor can / does refine.

Implements sgpp::datadriven::MultiGridRefinementFunctor.

References grids.

◆ getRefinementsNum()

size_t sgpp::datadriven::ClassificationRefinementFunctor::getRefinementsNum ( ) const

overridevirtual

Returns the maximal number of points that should be refined.

The maximal number of points to refine is set in the constructor of the implementing class.

Returns: number of points that should refined. Default value: 1.

Reimplemented from sgpp::base::RefinementFunctor.

References refinements_num.

◆ getRefinementThreshold()

double sgpp::datadriven::ClassificationRefinementFunctor::getRefinementThreshold ( ) const

overridevirtual

Returns the threshold for refinement.

Only the grid points with absolute value of refinement criterion (e.g., alpha or error) greater or equal to this threshold will be refined.

Returns: threshold value for refinement. Default value: 0.

Implements sgpp::base::RefinementFunctor.

References refinementThreshold.

◆ insertCoarseningCandidate()

void sgpp::datadriven::ClassificationRefinementFunctor::insertCoarseningCandidate	(	size_t	y,
		std::vector< std::multimap< double, std::tuple< size_t, size_t, bool > > > &	classMapsCoarsening,
		size_t	leafSeqNumber,
		double	score,
		std::tuple< size_t, size_t, bool >	candidate
	)

protected

Checks if a refinement candidate should be inserted into the map of candidates.

Referenced by adaptAllGrids().

◆ operator()()

double sgpp::datadriven::ClassificationRefinementFunctor::operator()	(	base::GridStorage &	storage,
		size_t	seq
	)		const

overridevirtual

This should be returning a refinement value for every grid point.

The point with the highest value will be refined first.

Parameters

storage	reference to the grids storage object
seq	sequence number in the coefficients array

Returns: refinement value

Implements sgpp::base::RefinementFunctor.

◆ preComputeEvaluations()

void sgpp::datadriven::ClassificationRefinementFunctor::preComputeEvaluations ( )

overridevirtual

Used if expensive computations (eg.

grid evaluations) are cached, usually for one refinement step.

Reimplemented from sgpp::datadriven::MultiGridRefinementFunctor.

◆ setGridIndex()

void sgpp::datadriven::ClassificationRefinementFunctor::setGridIndex ( size_t grid_index )

overridevirtual

Sets the index (into the vector of grids) of the grid to be refined.

Parameters

grid_index The index of the grid to be refined

Implements sgpp::datadriven::MultiGridRefinementFunctor.

◆ sgn()

int sgpp::datadriven::ClassificationRefinementFunctor::sgn ( double d ) const

protected

◆ start()

double sgpp::datadriven::ClassificationRefinementFunctor::start ( ) const

overridevirtual

Returns the lower bound of refinement criterion (e.g., alpha or error) (lower bound).

The refinement value of grid points to be refined have to be larger than this value

Returns: lower bound

Implements sgpp::base::RefinementFunctor.

◆ stepDown()

void sgpp::datadriven::ClassificationRefinementFunctor::stepDown	(	size_t	d,
		size_t	minDim,
		base::HashGridPoint &	gp,
		std::vector< std::pair< base::HashGridPoint, base::HashGridPoint > > &	neighbors
	)

protected

Iterate through the sparse grid tree.

References collectNeighbors(), sgpp::base::HashGridPoint::get(), sgpp::base::HashGridStorage::isContaining(), level, sgpp::base::HashGridPoint::set(), stepDown(), and total_grid.

Referenced by adaptAllGrids(), and stepDown().

Member Data Documentation

◆ alphas

std::vector<base::DataVector*> sgpp::datadriven::ClassificationRefinementFunctor::alphas

protected

Referenced by adaptAllGrids().

◆ coarsenings_num

size_t sgpp::datadriven::ClassificationRefinementFunctor::coarsenings_num

protected

Maximum number of coarsened points per iteration.

Referenced by adaptAllGrids().

◆ coarseningThreshold

double sgpp::datadriven::ClassificationRefinementFunctor::coarseningThreshold

protected

Referenced by adaptAllGrids().

◆ coarsenInitialPoints

bool sgpp::datadriven::ClassificationRefinementFunctor::coarsenInitialPoints

protected

Whether or not to coarsen initial grid points.

Referenced by adaptAllGrids().

◆ grids

std::vector<base::Grid*> sgpp::datadriven::ClassificationRefinementFunctor::grids

protected

Referenced by adaptAllGrids(), and getNumGrids().

◆ level_penalize

bool sgpp::datadriven::ClassificationRefinementFunctor::level_penalize

protected

◆ minimumCoarseningIndex

size_t sgpp::datadriven::ClassificationRefinementFunctor::minimumCoarseningIndex

protected

Referenced by adaptAllGrids().

◆ neighborRels

std::vector<std::tuple<size_t, size_t, size_t, bool, base::GridPoint::level_type> > sgpp::datadriven::ClassificationRefinementFunctor::neighborRels

protected

Referenced by adaptAllGrids(), and collectNeighbors().

◆ pre_comp_evals

std::vector<std::map<std::string, double> > sgpp::datadriven::ClassificationRefinementFunctor::pre_comp_evals

protected

Stores grid evaluations at all grids (vector) at the union of grid point coordinates over all grids (hashed by string representation in the map)

◆ priors

std::vector<double> sgpp::datadriven::ClassificationRefinementFunctor::priors

protected

Referenced by adaptAllGrids().

◆ refinements_num

size_t sgpp::datadriven::ClassificationRefinementFunctor::refinements_num

protected

Referenced by adaptAllGrids(), and getRefinementsNum().

◆ refinementThreshold

double sgpp::datadriven::ClassificationRefinementFunctor::refinementThreshold

protected

Referenced by adaptAllGrids(), and getRefinementThreshold().

◆ thresholdType

sgpp::base::AdaptivityThresholdType sgpp::datadriven::ClassificationRefinementFunctor::thresholdType

protected

Referenced by adaptAllGrids().

◆ total_grid

base::GridStorage sgpp::datadriven::ClassificationRefinementFunctor::total_grid

protected

Maximum index of the initial grid points.

Referenced by adaptAllGrids(), collectNeighbors(), and stepDown().

The documentation for this class was generated from the following files:

datadriven/src/sgpp/datadriven/functors/classification/ClassificationRefinementFunctor.hpp
datadriven/src/sgpp/datadriven/functors/classification/ClassificationRefinementFunctor.cpp

Public Member Functions

Protected Member Functions

Protected Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ ClassificationRefinementFunctor()

◆ ~ClassificationRefinementFunctor()

Member Function Documentation

◆ adaptAllGrids()

◆ collectNeighbors()

◆ getEvalVector()

◆ getNumGrids()

◆ getRefinementsNum()

◆ getRefinementThreshold()

◆ insertCoarseningCandidate()

◆ operator()()

◆ preComputeEvaluations()

◆ setGridIndex()

◆ sgn()

◆ start()

◆ stepDown()

Member Data Documentation

◆ alphas

◆ coarsenings_num

◆ coarseningThreshold

◆ coarsenInitialPoints

◆ grids

◆ level_penalize

◆ minimumCoarseningIndex

◆ neighborRels

◆ pre_comp_evals

◆ priors

◆ refinements_num

◆ refinementThreshold

◆ thresholdType

◆ total_grid