SG++
sgpp::datadriven::OperationInverseRosenblattTransformationKDE Class Reference

Do inverse transformation in all dimensions. More...

#include <OperationInverseRosenblattTransformationKDE.hpp>

Public Member Functions

virtual void doShuffledTransformation (base::DataMatrix &pointsUniform, base::DataMatrix &pointsCdf)
 
virtual void doTransformation (base::DataMatrix &pointsUniform, base::DataMatrix &pointsCdf)
 inverse Rosenblatt Transformation with mixed starting dimension More...
 
double doTransformation1D (double y, base::DataVector &samples1d, double sigma, double xlower, double xupper, double ylower, double yupper, base::DataVector &kern)
 do the inverse Rosenblatt transformation for one data point for given samples More...
 
double getMaxInversionError ()
 get the maximum error made during inversion More...
 
 OperationInverseRosenblattTransformationKDE (datadriven::GaussianKDE &kde, double sigmaFactor=6.0, double inversionEpsilon=1e-10)
 
virtual ~OperationInverseRosenblattTransformationKDE ()
 

Detailed Description

Do inverse transformation in all dimensions.

Constructor & Destructor Documentation

sgpp::datadriven::OperationInverseRosenblattTransformationKDE::OperationInverseRosenblattTransformationKDE ( datadriven::GaussianKDE kde,
double  sigmaFactor = 6.0,
double  inversionEpsilon = 1e-10 
)
sgpp::datadriven::OperationInverseRosenblattTransformationKDE::~OperationInverseRosenblattTransformationKDE ( )
virtual

Member Function Documentation

void sgpp::datadriven::OperationInverseRosenblattTransformationKDE::doTransformation ( base::DataMatrix pointsUniform,
base::DataMatrix pointsCdf 
)
virtual

inverse Rosenblatt Transformation with mixed starting dimension

Parameters
pointsUniformdata points to be transformed DataMatrix (rows: # of samples, columns: # of dims)
pointsCdfOutput DataMatrix (rows: # of samples, columns: # of dims)

References doTransformation1D(), sgpp::base::DataVector::get(), sgpp::base::DataMatrix::get(), sgpp::base::DataMatrix::getNrows(), sgpp::base::DataMatrix::getRow(), sgpp::datadriven::GaussianKDE::getSamples(), sgpp::base::DataVector::setAll(), and sgpp::base::DataMatrix::setRow().

double sgpp::datadriven::OperationInverseRosenblattTransformationKDE::doTransformation1D ( double  y,
base::DataVector samples1d,
double  sigma,
double  xlower,
double  xupper,
double  ylower,
double  yupper,
base::DataVector kern 
)

do the inverse Rosenblatt transformation for one data point for given samples

Parameters
ydata point to be inverted
samples1dtraining samples in the dimension to be transformed
sigmabandwidth of the kernels in the current dimension
xlowerlower bound for x-space
xupperupper bound for x-space
ylowerlower bound for y-space
yupperupper bound for y-space
kernkernel evaluations
Returns
error of inversion

References sgpp::base::DataVector::getSize(), M_1_SQRT2PI, and M_SQRT2.

Referenced by doShuffledTransformation(), and doTransformation().


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