sgpp::datadriven::SystemMatrixLeastSquaresIdentity Class Reference

Class that implements the virtual class base::OperationMatrix for the application of classification for the Systemmatrix. More...

#include <SystemMatrixLeastSquaresIdentity.hpp>

Inheritance diagram for sgpp::datadriven::SystemMatrixLeastSquaresIdentity:

Public Member Functions
virtual void	generateb (base::DataVector &classes, base::DataVector &b)
	Generates the right hand side of the classification equation.
virtual void	mult (base::DataVector &alpha, base::DataVector &result)
	starts the Multiplication with the matrix
virtual void	prepareGrid ()
	forward declaration
void	setImplementation (datadriven::OperationMultipleEvalConfiguration operationConfiguration)
	SystemMatrixLeastSquaresIdentity (base::Grid &SparseGrid, base::DataMatrix &trainData, double lambda)
	Std-Constructor.
virtual	~SystemMatrixLeastSquaresIdentity ()
	Std-Destructor.
Public Member Functions inherited from sgpp::datadriven::DMSystemMatrixBase
	DMSystemMatrixBase (base::DataMatrix &trainData, double lambda)
	Std-Constructor.
virtual void	getTimers (double &timeMult, double &computeMult, double &timeMultTrans, double &computeMultTrans)
	gets the timer's values by saving them into call by reference values
virtual void	resetTimers ()
	resets all timers to 0
virtual	~DMSystemMatrixBase ()
	Std-Destructor.
Public Member Functions inherited from sgpp::base::OperationMatrix
	OperationMatrix ()
	Constructor.
virtual	~OperationMatrix ()
	Destructor.

Additional Inherited Members
Protected Attributes inherited from sgpp::datadriven::DMSystemMatrixBase
double	completeTimeMult_
	time needed for Mult
double	completeTimeMultTrans_
	time needed for Mult transposed
double	computeTimeMult_
	time needed only for the computation of mult, interesting on accelerator boards
double	computeTimeMultTrans_
	time needed only for the computation of mult transposed, interesting on accelerator boards
base::DataMatrix	dataset_
	the dataset
double	lambda_
	the lambda, the regularisation parameter
base::SGppStopwatch *	myTimer_
	Stopwatch needed to determine the durations of mult and mult transposed.

Detailed Description

Class that implements the virtual class base::OperationMatrix for the application of classification for the Systemmatrix.

The Identity matrix is used as regularization operator.

For the Operation B's mult and mutlTransposed functions vectorized formulations are used.

Constructor & Destructor Documentation

SystemMatrixLeastSquaresIdentity()

sgpp::datadriven::SystemMatrixLeastSquaresIdentity::SystemMatrixLeastSquaresIdentity	(	base::Grid &	SparseGrid,
		base::DataMatrix &	trainData,
		double	lambda
	)

Std-Constructor.

Parameters

SparseGrid	reference to the sparse grid
trainData	reference to base::DataMatrix that contains the training data
lambda	the lambda, the regression parameter

References sgpp::op_factory::createOperationMultipleEval(), sgpp::datadriven::DMSystemMatrixBase::dataset_, and sgpp::base::DataMatrix::getNrows().

~SystemMatrixLeastSquaresIdentity()

sgpp::datadriven::SystemMatrixLeastSquaresIdentity::~SystemMatrixLeastSquaresIdentity ( )

virtual

Std-Destructor.

Member Function Documentation

generateb()

void sgpp::datadriven::SystemMatrixLeastSquaresIdentity::generateb ( base::DataVector &  classes, base::DataVector &  b  )

virtual

Generates the right hand side of the classification equation.

Parameters

classes	the class information of the training data
b	reference to the vector that will contain the result of the matrix vector multiplication on the rhs

Implements sgpp::datadriven::DMSystemMatrixBase.

References sgpp::datadriven::DMSystemMatrixBase::completeTimeMultTrans_, sgpp::datadriven::DMSystemMatrixBase::computeTimeMultTrans_, sgpp::datadriven::DMSystemMatrixBase::myTimer_, sgpp::base::SGppStopwatch::start(), and sgpp::base::SGppStopwatch::stop().

mult()

void sgpp::datadriven::SystemMatrixLeastSquaresIdentity::mult ( base::DataVector &  alpha, base::DataVector &  result  )

virtual

starts the Multiplication with the matrix

Parameters

alpha	DataVector that contains the ansatzfunctions' coefficients
result	DataVector into which the result of the Laplace operation is stored

Implements sgpp::datadriven::DMSystemMatrixBase.

References alpha, sgpp::base::DataVector::axpy(), sgpp::datadriven::DMSystemMatrixBase::completeTimeMult_, sgpp::datadriven::DMSystemMatrixBase::completeTimeMultTrans_, sgpp::datadriven::DMSystemMatrixBase::computeTimeMult_, sgpp::datadriven::DMSystemMatrixBase::computeTimeMultTrans_, sgpp::datadriven::DMSystemMatrixBase::lambda_, sgpp::datadriven::DMSystemMatrixBase::myTimer_, sgpp::base::SGppStopwatch::start(), and sgpp::base::SGppStopwatch::stop().

prepareGrid()

void sgpp::datadriven::SystemMatrixLeastSquaresIdentity::prepareGrid ( )

virtual

forward declaration

rebuilds the base::DataMatrix for Level and Index this routine is needed for supporting adaptive grids with vectorized high performance kernels

Reimplemented from sgpp::datadriven::DMSystemMatrixBase.

setImplementation()

void sgpp::datadriven::SystemMatrixLeastSquaresIdentity::setImplementation

(

datadriven::OperationMultipleEvalConfiguration

operationConfiguration

)

References sgpp::op_factory::createOperationMultipleEval(), and sgpp::datadriven::DMSystemMatrixBase::dataset_.

---

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

• datadriven/src/sgpp/datadriven/algorithm/SystemMatrixLeastSquaresIdentity.hpp
• datadriven/src/sgpp/datadriven/algorithm/SystemMatrixLeastSquaresIdentity.cpp