sgpp::datadriven::LearnerBaseSP Class Reference

Abstract class that implements a regression/classification learner based on spatial adaptive Sparse Grids. More...

#include <LearnerBaseSP.hpp>

Public Member Functions
void	dumpFunction (std::string tFilename, size_t resolution)
	simple dump of sparse grid function into file, e.g.
void	dumpGrid (std::string tFilename)
	simple dump of grid points into file, e.g.
virtual double	getAccuracy (const sgpp::base::DataVectorSP &classesComputed, const sgpp::base::DataVectorSP &classesReference, const float threshold=0.0)
	compute the accuracy for given testDataset.
virtual double	getAccuracy (sgpp::base::DataMatrixSP &testDataset, const sgpp::base::DataVectorSP &classesReference, const float threshold=0.0)
	compute the accuracy for given testDataset.
virtual ClassificatorQuality	getCassificatorQuality (const sgpp::base::DataVectorSP &classesComputed, const sgpp::base::DataVectorSP &classesReference, const float threshold=0.0)
	compute the quality for given testDataset, classification ONLY!
virtual ClassificatorQuality	getCassificatorQuality (sgpp::base::DataMatrixSP &testDataset, const sgpp::base::DataVectorSP &classesReference, const float threshold=0.0)
	compute the quality for given testDataset, classification ONLY! test is automatically called in order to determine the regression values of the current learner
bool	getIsRegression () const
	determines the current mode
bool	getIsVerbose () const
	determines the current verbose mode of learner
	LearnerBaseSP (const bool isRegression, const bool isVerbose=true)
	Constructor.
	LearnerBaseSP (const LearnerBaseSP &copyMe)
	Copy-Constructor.
	LearnerBaseSP (std::string tGridFilename, std::string tAlphaFilename, const bool isRegression, const bool isVerbose=true)
	Constructor.
virtual sgpp::base::DataVectorSP	predict (sgpp::base::DataMatrixSP &testDataset)
	executes a Regression test for a given dataset and returns the result
void	setIsVerbose (const bool isVerbose)
	sets the current verbose mode of learner
void	store (std::string tGridFilename, std::string tAlphaFilename)
	store the grid and its current coefficients into files for further usage.
LearnerTiming	train (sgpp::base::DataMatrixSP &testDataset, sgpp::base::DataVectorSP &classes, const sgpp::base::RegularGridConfiguration &gridConfig, const sgpp::solver::SLESolverSPConfiguration &SolverConfig, const float lambdaRegularization)
	Learning a dataset with regular sparse grids.
virtual LearnerTiming	train (sgpp::base::DataMatrixSP &testDataset, sgpp::base::DataVectorSP &classes, const sgpp::base::RegularGridConfiguration &gridConfig, const sgpp::solver::SLESolverSPConfiguration &SolverConfigRefine, const sgpp::solver::SLESolverSPConfiguration &SolverConfigFinal, const sgpp::base::AdaptivityConfiguration &adaptivityConfig, bool testAccDuringAdapt, const float lambdaRegularization)
	Learning a dataset with spatially adaptive sparse grids.
virtual	~LearnerBaseSP ()
	Destructor.

Protected Member Functions
virtual sgpp::datadriven::DMSystemMatrixBaseSP *	createDMSystem (sgpp::base::DataMatrixSP &trainDataset, float lambdaRegularization)=0
	abstract method that constructs the corresponding system of linear equations Derived classes MUST overwrite this functions!
virtual void	InitializeGrid (const sgpp::base::RegularGridConfiguration &gridConfig)
	Initialize the grid and its coefficients.
virtual void	postProcessing (const sgpp::base::DataMatrixSP &trainDataset, const sgpp::solver::SLESolverType &solver, const size_t numNeededIterations)
	Hook-Method for post-processing after each refinement learning.
virtual void	preProcessing ()
	Hook-Method for pre-processing before starting learning.

Protected Attributes
sgpp::base::DataVectorSP *	alpha_
	the grid's coefficients
double	execTime_
	execution time
double	GByte_
	number of transferred Gbytes
double	GFlop_
	number of executed Floating Point operations
sgpp::base::Grid *	grid_
	sparse grid object
bool	isRegression_
	is regression selected
bool	isTrained_
	is the grid trained
bool	isVerbose_
	is verbose output enabled

Detailed Description

Abstract class that implements a regression/classification learner based on spatial adaptive Sparse Grids.

Furthermore this class is intended to provide a framework for adavanded regression and classification methods by allowing to override basic routines like train and test.

This versions supports single precision datatypes.

Constructor & Destructor Documentation

LearnerBaseSP() [1/3]

sgpp::datadriven::LearnerBaseSP::LearnerBaseSP ( const bool  isRegression, const bool  isVerbose = true  )

explicit

Constructor.

Parameters

isRegression	flag for regression
isVerbose	flag for verbose output

LearnerBaseSP() [2/3]

sgpp::datadriven::LearnerBaseSP::LearnerBaseSP	(	std::string	tGridFilename,
		std::string	tAlphaFilename,
		const bool	isRegression,
		const bool	isVerbose = true
	)

Constructor.

Parameters

tGridFilename	path to file that contains a serialized grid
tAlphaFilename	path to file that contains the grid's coefficients
isRegression	set to true if a regression task should be executed
isVerbose	set to true in order to allow console output

LearnerBaseSP() [3/3]

sgpp::datadriven::LearnerBaseSP::LearnerBaseSP

(

const LearnerBaseSP &

copyMe

)

Copy-Constructor.

Parameters

copyMe

LearnerBase that should be duplicated

References alpha_, execTime_, GByte_, GFlop_, grid_, isRegression_, isTrained_, sgpp::base::Grid::serialize(), and sgpp::base::Grid::unserialize().

~LearnerBaseSP()

sgpp::datadriven::LearnerBaseSP::~LearnerBaseSP ( )

virtual

Destructor.

References alpha_, and grid_.

Member Function Documentation

createDMSystem()

virtual sgpp::datadriven::DMSystemMatrixBaseSP * sgpp::datadriven::LearnerBaseSP::createDMSystem ( sgpp::base::DataMatrixSP &  trainDataset, float  lambdaRegularization  )

protectedpure virtual

abstract method that constructs the corresponding system of linear equations Derived classes MUST overwrite this functions!

Parameters

trainDataset	training dataset
lambdaRegularization	lambda regularization parameter

Referenced by train().

dumpFunction()

void sgpp::datadriven::LearnerBaseSP::dumpFunction	(	std::string	tFilename,
		size_t	resolution
	)

simple dump of sparse grid function into file, e.g.

used to plot with gnuplot.

only executed if grid is trained and number of dimensions <= 2.

Parameters

tFilename	filename to store the dump to
resolution	resolution of function plot

References alpha_, sgpp::base::PrecisionConverter::convertDataVectorSPToDataVector(), sgpp::base::Grid::getDimension(), sgpp::base::DataVectorSP::getSize(), grid_, isTrained_, and sgpp::base::GridPrinter::printGrid().

dumpGrid()

void sgpp::datadriven::LearnerBaseSP::dumpGrid

(

std::string

tFilename

)

simple dump of grid points into file, e.g.

used to plot with gnuplot

only executed if grid is trained

Parameters

tFilename

filename to store the dump to

References alpha_, sgpp::base::PrecisionConverter::convertDataVectorSPToDataVector(), sgpp::base::DataVectorSP::getSize(), grid_, isTrained_, and sgpp::base::GridPrinter::printSparseGrid().

getAccuracy() [1/2]

double sgpp::datadriven::LearnerBaseSP::getAccuracy ( const sgpp::base::DataVectorSP &  classesComputed, const sgpp::base::DataVectorSP &  classesReference, const float  threshold = 0.0  )

virtual

compute the accuracy for given testDataset.

In case if classification (isRegression == false) this routine returns the learner's accuracy In case of regressions (isRegression == true) this routine returns the learner's MSE

Parameters

classesComputed	regression results of the test dataset
classesReference	reference labels of the test dataset
threshold	threshold used for classification, ignored when performing regressions

Returns

accuracy, percent or MSE, depending on the execution mode

References sgpp::base::DataVectorSP::get(), sgpp::base::DataVectorSP::getSize(), and isRegression_.

getAccuracy() [2/2]

double sgpp::datadriven::LearnerBaseSP::getAccuracy ( sgpp::base::DataMatrixSP &  testDataset, const sgpp::base::DataVectorSP &  classesReference, const float  threshold = 0.0  )

virtual

compute the accuracy for given testDataset.

test is automatically called in order to determine the regression values of the current learner

In case if classification (isRegression == false) this routine returns the learner's accuracy In case of regressions (isRegression == true) this routine returns the learner's MSE

Parameters

testDataset	dataset to be tested
classesReference	reference labels of the test dataset
threshold	threshold used for classification, ignored when performing regressions

Returns

accuracy, percent or MSE, depending on the execution mode

References getAccuracy(), and predict().

Referenced by getAccuracy(), and train().

getCassificatorQuality() [1/2]

ClassificatorQuality sgpp::datadriven::LearnerBaseSP::getCassificatorQuality ( const sgpp::base::DataVectorSP &  classesComputed, const sgpp::base::DataVectorSP &  classesReference, const float  threshold = 0.0  )

virtual

compute the quality for given testDataset, classification ONLY!

Parameters

classesComputed	regression results of the test dataset
classesReference	reference labels of the test dataset
threshold	threshold used for classification, ignored when performing regressions

Returns

quality structure containing tp, tn, fp, fn counts

References sgpp::datadriven::ClassificatorQuality::falseNegative_, sgpp::datadriven::ClassificatorQuality::falsePositive_, sgpp::base::DataVectorSP::get(), sgpp::base::DataVectorSP::getSize(), isRegression_, sgpp::datadriven::ClassificatorQuality::trueNegative_, and sgpp::datadriven::ClassificatorQuality::truePositive_.

getCassificatorQuality() [2/2]

ClassificatorQuality sgpp::datadriven::LearnerBaseSP::getCassificatorQuality ( sgpp::base::DataMatrixSP &  testDataset, const sgpp::base::DataVectorSP &  classesReference, const float  threshold = 0.0  )

virtual

compute the quality for given testDataset, classification ONLY! test is automatically called in order to determine the regression values of the current learner

Parameters

testDataset	dataset to be tested
classesReference	reference labels of the test dataset
threshold	threshold used for classification, ignored when performing regressions

Returns

quality structure containing tp, tn, fp, fn counts

References getCassificatorQuality(), and predict().

Referenced by getCassificatorQuality().

getIsRegression()

bool sgpp::datadriven::LearnerBaseSP::getIsRegression

(

)

const

determines the current mode

Returns

returns whether the current mode is regression or not

References isRegression_.

getIsVerbose()

bool sgpp::datadriven::LearnerBaseSP::getIsVerbose

(

)

const

determines the current verbose mode of learner

Returns

returns whether the current learner has verbose output

References isVerbose_.

InitializeGrid()

void sgpp::datadriven::LearnerBaseSP::InitializeGrid ( const sgpp::base::RegularGridConfiguration &  gridConfig )

protectedvirtual

Initialize the grid and its coefficients.

Parameters

gridConfig

structure which describes the regular start grid

References alpha_, sgpp::base::GeneralGridConfiguration::dim_, sgpp::base::Grid::getGenerator(), sgpp::base::Grid::getSize(), grid_, sgpp::base::GeneralGridConfiguration::level_, sgpp::base::Linear, sgpp::base::LinearBoundary, sgpp::base::ModLinear, sgpp::base::GridGenerator::regular(), sgpp::base::DataVectorSP::setAll(), and sgpp::base::GeneralGridConfiguration::type_.

Referenced by train().

postProcessing()

void sgpp::datadriven::LearnerBaseSP::postProcessing ( const sgpp::base::DataMatrixSP &  trainDataset, const sgpp::solver::SLESolverType &  solver, const size_t  numNeededIterations  )

protectedvirtual

Hook-Method for post-processing after each refinement learning.

can be overwritten by derived classes

Parameters

trainDataset	matrix with training data set
solver	solver
numNeededIterations	number of required iterations

References execTime_, and isVerbose_.

Referenced by train().

predict()

sgpp::base::DataVectorSP sgpp::datadriven::LearnerBaseSP::predict ( sgpp::base::DataMatrixSP &  testDataset )

virtual

executes a Regression test for a given dataset and returns the result

Parameters

testDataset

dataset that is evaluated with the current learner

Returns

regression values of testDataset

References alpha_, sgpp::base::PrecisionConverter::convertDataMatrixSPToDataMatrix(), sgpp::base::PrecisionConverter::convertDataVectorSPToDataVector(), sgpp::base::PrecisionConverter::convertDataVectorToDataVectorSP(), sgpp::op_factory::createOperationMultipleEval(), sgpp::base::DataMatrixSP::getNcols(), sgpp::base::DataMatrixSP::getNrows(), sgpp::base::Grid::getSize(), grid_, and sgpp::base::OperationMultipleEval::mult().

Referenced by getAccuracy(), and getCassificatorQuality().

preProcessing()

void sgpp::datadriven::LearnerBaseSP::preProcessing ( )

protectedvirtual

Hook-Method for pre-processing before starting learning.

can be overwritten by derived classes

Referenced by train().

setIsVerbose()

void sgpp::datadriven::LearnerBaseSP::setIsVerbose

(

const bool

isVerbose

)

sets the current verbose mode of learner

Parameters

isVerbose

the current learner's verbose output

References isVerbose_.

store()

void sgpp::datadriven::LearnerBaseSP::store	(	std::string	tGridFilename,
		std::string	tAlphaFilename
	)

store the grid and its current coefficients into files for further usage.

Parameters

tGridFilename	filename of grid file
tAlphaFilename	filename of coefficient file

train() [1/2]

LearnerTiming sgpp::datadriven::LearnerBaseSP::train	(	sgpp::base::DataMatrixSP &	testDataset,
		sgpp::base::DataVectorSP &	classes,
		const sgpp::base::RegularGridConfiguration &	gridConfig,
		const sgpp::solver::SLESolverSPConfiguration &	SolverConfig,
		const float	lambdaRegularization
	)

Learning a dataset with regular sparse grids.

Parameters

testDataset	the training dataset
classes	classes corresponding to the training dataset
gridConfig	configuration of the regular grid
SolverConfig	configuration of the SLE solver
lambdaRegularization	regularization parameter lambda

References adaptivityConfig, sgpp::base::AdaptivityConfiguration::maxLevelType_, sgpp::base::AdaptivityConfiguration::numRefinementPoints_, sgpp::base::AdaptivityConfiguration::numRefinements_, sgpp::base::AdaptivityConfiguration::percent_, sgpp::base::AdaptivityConfiguration::refinementThreshold_, and train().

train() [2/2]

LearnerTiming sgpp::datadriven::LearnerBaseSP::train ( sgpp::base::DataMatrixSP &  testDataset, sgpp::base::DataVectorSP &  classes, const sgpp::base::RegularGridConfiguration &  gridConfig, const sgpp::solver::SLESolverSPConfiguration &  SolverConfigRefine, const sgpp::solver::SLESolverSPConfiguration &  SolverConfigFinal, const sgpp::base::AdaptivityConfiguration &  adaptivityConfig, bool  testAccDuringAdapt, const float  lambdaRegularization  )

virtual

Learning a dataset with spatially adaptive sparse grids.

Parameters

testDataset	the training dataset
classes	classes corresponding to the training dataset
gridConfig	configuration of the regular start grid
SolverConfigRefine	configuration of the SLE solver during the adaptive refinements of the grid
SolverConfigFinal	configuration of the final SLE solving step on the refined grid
adaptivityConfig	configuration of the adaptivity strategy
testAccDuringAdapt	set to true if the training accuracy should be determined in evert refinement step
lambdaRegularization	regularization parameter lambda

References adaptivityConfig, alpha_, sgpp::solver::BiCGSTAB, sgpp::solver::CG, sgpp::base::PrecisionConverter::convertDataVectorSPToDataVector(), createDMSystem(), sgpp::solver::SLESolverSPConfiguration::eps_, execTime_, GByte_, sgpp::datadriven::LearnerTiming::GByte_, sgpp::datadriven::DMSystemMatrixBaseSP::generateb(), getAccuracy(), sgpp::base::Grid::getGenerator(), sgpp::base::DataMatrixSP::getNrows(), sgpp::solver::SGSolverSP::getNumberIterations(), sgpp::solver::SGSolverSP::getResiduum(), sgpp::base::DataVectorSP::getSize(), sgpp::base::Grid::getSize(), sgpp::datadriven::DMSystemMatrixBaseSP::getTimers(), GFlop_, sgpp::datadriven::LearnerTiming::GFlop_, grid_, InitializeGrid(), isRegression_, isTrained_, isVerbose_, sgpp::solver::SLESolverSPConfiguration::maxIterations_, sgpp::base::AdaptivityConfiguration::numRefinementPoints_, sgpp::base::AdaptivityConfiguration::numRefinements_, postProcessing(), preProcessing(), sgpp::datadriven::DMSystemMatrixBaseSP::rebuildLevelAndIndex(), sgpp::base::GridGenerator::refine(), sgpp::base::AdaptivityConfiguration::refinementThreshold_, sgpp::base::DataVectorSP::resizeZero(), sgpp::solver::SGSolverSP::setEpsilon(), sgpp::solver::SGSolverSP::setMaxIterations(), sgpp::solver::SLESolverSP::solve(), sgpp::base::SGppStopwatch::start(), sgpp::base::SGppStopwatch::stop(), sgpp::datadriven::LearnerTiming::timeComplete_, sgpp::datadriven::LearnerTiming::timeMultComplete_, sgpp::datadriven::LearnerTiming::timeMultCompute_, sgpp::datadriven::LearnerTiming::timeMultTransComplete_, sgpp::datadriven::LearnerTiming::timeMultTransCompute_, sgpp::datadriven::LearnerTiming::timeRegularization_, and sgpp::solver::SLESolverSPConfiguration::type_.

Referenced by train().

Member Data Documentation

alpha_

sgpp::base::DataVectorSP* sgpp::datadriven::LearnerBaseSP::alpha_

protected

the grid's coefficients

Referenced by dumpFunction(), dumpGrid(), InitializeGrid(), LearnerBaseSP(), predict(), train(), and ~LearnerBaseSP().

execTime_

double sgpp::datadriven::LearnerBaseSP::execTime_

protected

execution time

Referenced by LearnerBaseSP(), postProcessing(), and train().

GByte_

double sgpp::datadriven::LearnerBaseSP::GByte_

protected

number of transferred Gbytes

Referenced by LearnerBaseSP(), and train().

GFlop_

double sgpp::datadriven::LearnerBaseSP::GFlop_

protected

number of executed Floating Point operations

Referenced by LearnerBaseSP(), and train().

grid_

sgpp::base::Grid* sgpp::datadriven::LearnerBaseSP::grid_

protected

sparse grid object

Referenced by dumpFunction(), dumpGrid(), InitializeGrid(), LearnerBaseSP(), predict(), train(), and ~LearnerBaseSP().

isRegression_

bool sgpp::datadriven::LearnerBaseSP::isRegression_

protected

is regression selected

Referenced by getAccuracy(), getCassificatorQuality(), getIsRegression(), LearnerBaseSP(), and train().

isTrained_

bool sgpp::datadriven::LearnerBaseSP::isTrained_

protected

is the grid trained

Referenced by dumpFunction(), dumpGrid(), LearnerBaseSP(), and train().

isVerbose_

bool sgpp::datadriven::LearnerBaseSP::isVerbose_

protected

is verbose output enabled

Referenced by getIsVerbose(), postProcessing(), setIsVerbose(), and train().

---

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

• datadriven/src/sgpp/datadriven/application/LearnerBaseSP.hpp
• datadriven/src/sgpp/datadriven/application/LearnerBaseSP.cpp