Description of the class TKriging. More...

#include <TKriging.h>

Inheritance diagram for URANIE::Modeler::TKriging:

Collaboration diagram for URANIE::Modeler::TKriging:

[legend]

Public Member Functions
Constructor and Destructor
	TKriging (Int_t nX, Int_t nS, Double_t xObs, Double_t xNormParams, Double_t yObs, Double_t gamma, Double_t iL, Double_t gpVar, Double_t vErrMes, URANIE::Modeler::TCorrelationFunction corrFunc, TString varNames, Int_t nP=0, Double_t hbeta=NULL, Double_t vbeta=NULL, Double_t M=NULL, Double_t iR=NULL, TString trend="")
	Default constructor.

void	initMatrices ()
	Initialisation of the matrices.

virtual	~TKriging ()
	Default destructor.

Get methods
const char *	getInputNames ()
	Return the input variable names.

const char *	getOutputName ()
	Return the output variable name.

Bool_t	hasTrend ()
	Return kTRUE if a deterministic trend has been defined.

TCorrelationFunction *	getCorrFunction ()
	Return a pointer to the correlation function of the gaussian process.

Int_t	getNInputs ()
	Return the number of input parameters.

Int_t	getNObs ()
	Return the number of observations used to build the GP model.

Int_t	getNCorrFunctionParams ()
	Return the number of parameters of the correlation function.

Int_t	getNCorrLengths ()
	Return the number of correlation lengths.

Int_t	getNTrendParams ()
	Return the number of parameters of the deterministic trend.

Double_t *	getObsInputs ()
	Return the observations inputs as a flat array: [a1, b1, c1, a2, b2, c2,...].

Double_t *	getObsOutputs ()
	Return the observations outputs.

Double_t *	getObsNormParams ()
	Return the normalisation parameters of the inputs as a flat array: [aMin, aMax, bMin, bMax,...].

Double_t *	getGPParams ()
	Return the array of the Gaussian Process parameters.

Double_t *	getTrendParams ()
	Return the array of the trend parameters.

Double_t *	getTrendParamsCovMatrix ()
	Return the array of the trend parameters covariance matrix.

TObjArray *	getTrendFormula ()
	Return the deterministic trend list of formulas.

TString	getTrendString ()
	Return the deterministic trend string.

Double_t	getVariance ()
	Return the variance of the gaussian process.

Double_t	getMeasurementErrorVariance ()
	Return the variance of the measurement error.

Double_t *	getiL ()
	Return the iL matrix as an array (cf. internal variable description)

Double_t *	getM ()
	Return the M matrix as an array (cf. internal variable description)

Double_t *	getiR ()
	Return the iR matrix as an array (cf. internal variable description)

Double_t *	getLooErrors ()
	Return the vector of the leave one out errors.

Double_t *	getLooVariances ()
	Return the vector of the leave one out prediction variances.

void	getLooErrors (double *arr, int size)
	Return the vector of the leave one out errors.

void	getLooVariances (double *arr, int size)
	Return the vector of the leave one out prediction variances.

Double_t	getLooRMSE ()
	Return the leave one out mean squared error.

Double_t	getLooQ2 ()
	Return the Q2 of the model on the observation.

URANIE::DataServer::TDataServer *	getLooData ()
	Return a pointer to a TDataServer containing the observations and the Leave One Out results.

Double_t *	getCorrLengthsNormalised ()
	Return the normalised correlation lengths.

Double_t *	getCorrLengths ()
	Return the correlation lengths.

Double_t *	getCorrFunctionParams ()
	Return all the parameters of the correlation function.

Evaluation
Int_t	eval (Double_t x0, Double_t y0, int=0)
	Evaluate the gaussian process at the new location x0.

void	estimate (URANIE::DataServer::TDataServer tdsQuery, TString listOfInputs="", TString listOfOutputs="", Option_t option="", Bool_t useGPU=false)

void	estimate_CPU (URANIE::DataServer::TDataServer tdsQuery, TString listOfInputs="", TString listOfOutputs="", Option_t option="")

void	estimateWithCov (URANIE::DataServer::TDataServer tdsQuery, TString listOfInputs="", TString listOfOutputs="", TString realValue="", Option_t option="")

Tools
Exporting the kriging model as a stand alone C function is difficult for various reasons. The main problem is that we would have to store potentially huge matrices as text, which is highly unpractical. It was thus decided to postpone this export as long as a workable solution isn't found.
void	exportFunction (const char lang, const char file, const char name, Option_t option="")

virtual void	exportModelCplusplus (std::ofstream sourcefile, const char name="")
	Export the model as a C function (not implemented)

virtual void	exportModelFortran (std::ofstream *sourcefile)
	Export the model as a Fortran function (not implemented)

virtual void	exportModelPMML (const char file, const char name, Option_t *option) const

void	freeze ()

Printing Log
virtual void	printLog (Option_t *option="")

Public Attributes
TString	_outputVarName
	Name of the output variable. This information can be used to remind the user of the original names of the variables.

TString	_inputVarNames
	Name of the inputs variables separated by ":". This information can be used to remind the user of the original names of the variables.

TObjArray *	_inputNamesList
	list of the inputs' names.

Int_t	_nX
	Number of input attributes

Int_t	_nS
	Number of observations.

Int_t	_nYDeclared
	Number of attribute added to the tkiging object before running it.

Int_t	_nXDeclared
	Number of attribute added to the tkiging object before running it.

Double_t *	_xObs
	The matrix of inputs (dimension: _nX * _nS)

Double_t *	_xNormParams
	The matrix of normalisation parameters. Contains min and max value of each input variable: [aMin, aMax, bMin, bMax,...] (size: _nX * 2)

Double_t *	_yObs
	Array representation of the outputs (size: _nS)

Double_t *	_iL
	Inverse of the matrix L, Cholesky decomposition of K (size: _nS * _nS)

Double_t *	_gamma
	Array of parameters of the gaussian process (size: _nS)

Double_t	_sigma2
	Variance of the gaussian process.

Double_t	_vErrMes
	variance of the measurement error

TCorrelationFunction *	_correlationFunction
	The correlation function.

Double_t *	_corrLengths
	The correlation lengths in the real input space.

Double_t *	_normCorrLengths
	The correlation lengths in the normalised input space. This corresponds to the correlation lengths stored in the correlation function.

Bool_t	bUse_normalisation
	if kFALSE, the matrix H will be computed with unormalised input data. This happens when the user is providing his own trend function. When the trend is constructed automatically from a trend type, this attribute is set to kTRUE.

Bool_t	bHas_trend
	kFALSE if no deterministic trend is defined, kTRUE otherwise

TString	_trendString
	Character string containing the trend parameters separated by ":" or a trend type ("const" or "linear")

TObjArray *	_trendCoefList
	List of the formulas of the trend coefficients (_nP elements)

Int_t	_nP
	Number of coefficients of the deterministic trend.

Double_t *	_hbeta
	Estimates of the deterministic trend parameters (size: _nP)

Double_t *	_vbeta
	Covariance matrix of the trend parameters (size: _nP*_nP)

Double_t *	_M
	Matrix M = L^{-T} H (size: _nS * _nP)

Double_t *	_iR
	Inverse of the matrix R, from the QR decomposition of M (size: _nP * _nP)

Double_t *	_errorLoo
	If available, vector of the Leave One Out errors for the model.

Double_t *	_varLoo
	If available, vector of the variances of the Leave One Out errors for the model.

Double_t	_mseLoo
	if available, Leave one Out mean squared error of the model.

Private Attributes
Bool_t	_blog
	Boolean for edit the log.

Set methods
void	setCorrFunction (TCorrelationFunction *corrFunc)
	Set the correlation function of the gaussian process.

void	setVariance (Double_t var)
	Set the variance of the gaussian process.

void	setLooErrors (Double_t err, Double_t var, Double_t mse)
	Set the information about the leave one out error of the model.

void	createTrend ()
	Create the list of trend coefficients.

void	computeCovarianceMatrix (int ns, double xObs, bool bVariance_in_factor, bool bHas_measurement_error, double alpha, double reg, double sigma2, double Vmes, double C, double K)

static void	StatMat (std::string title, int n1, int n2, Double_t *v)

static void	StatMat (std::string title, int n1, int n2, Double_t *v, std::vector< double > &tvmin, std::vector< double > &tvmax, std::vector< double > &tmean, std::vector< double > &tvar, std::vector< double > &tnorm)

static void	StatVect (std::string title, int n, Double_t *v)

static void	StatVect (std::string title, int n, Double_t *v, double &vmin, double &vmax, double &mean, double &var, double &norm)

Detailed Description

Description of the class TKriging.

This class defines a Kriging model. It can predict the output of new observations using the TKriging::eval function or the URANIE::Relauncher module formalism. It is created by TGPBuilder::buildGP. It is a wrapper to the gpPredX functions from the "gaussian process Library" (gpL), by Jean-Marc Martinez.

Constructor & Destructor Documentation

◆ TKriging()

URANIE::Modeler::TKriging::TKriging	(	Int_t	nX,
		Int_t	nS,
		Double_t *	xObs,
		Double_t *	xNormParams,
		Double_t *	yObs,
		Double_t *	gamma,
		Double_t *	iL,
		Double_t	gpVar,
		Double_t	vErrMes,
		URANIE::Modeler::TCorrelationFunction *	corrFunc,
		TString	varNames,
		Int_t	nP = `0`,
		Double_t *	hbeta = `NULL`,
		Double_t *	vbeta = `NULL`,
		Double_t *	M = `NULL`,
		Double_t *	iR = `NULL`,
		TString	trend = `""`
	)

Default constructor.

Constructs a TKriging from the raw matrices and other stuff produced by TGPBuilder. This constructor is not supposed to be used by a human being.

Parameters

nX	(Int_t): number of variables
nS	(Int_t): number of observations
xObs	(Double_t): the normalised observations (size = nX nS)
xNormParams	(Double_t): the normalisation parameters of xObs (size = 2 nX)
yObs	(Double_t*); the outputs (size = _nS)
gamma	(Double_t*): parameters of the Kriging model (size = nS)
iL	(Double_t): inverse of matrix L (size = nS nS)
gpVar	(Double_t): variance of the Gaussian Process
vErrMes	(Double_t): variance of the measurement error
corrFunc	(TCorrelationFunction*): correlation function of the Gaussian Process
varNames	(TString ): name of the variables separated by ":". The last name is the output variable.
nP	(Int_t): number of parameters of the trend (default = 0)
hbeta	(Double_t *): parameters of the trend (size = nP) (default = NULL)
M	(Double_t): matrice M (size = nS nP) (default = NULL)
iR	(Double_t): inverse de la matrice R (size = nP nP) (default = NULL)
trend	(TString): string listing the trend coefficients (default = "")

Note: the TKriging object is copying every information provided. This means that the source of this information can be safely destroyed after creating the TKriging object.

Referenced by ClassImp().

◆ ~TKriging()

virtual URANIE::Modeler::TKriging::~TKriging ( )

virtual

Default destructor.

Referenced by ClassImp().

Member Function Documentation

◆ computeCovarianceMatrix()

void URANIE::Modeler::TKriging::computeCovarianceMatrix	(	int	ns,
		double *	xObs,
		bool	bVariance_in_factor,
		bool	bHas_measurement_error,
		double	alpha,
		double	reg,
		double	sigma2,
		double *	Vmes,
		double *	C,
		double *	K
	)

Referenced by ClassImp().

◆ createTrend()

void URANIE::Modeler::TKriging::createTrend ( )

Create the list of trend coefficients.

Referenced by ClassImp().

◆ estimate()

void URANIE::Modeler::TKriging::estimate	(	URANIE::DataServer::TDataServer *	tdsQuery,
		TString	listOfInputs = `""`,
		TString	listOfOutputs = `""`,
		Option_t *	option = `""`,
		Bool_t	useGPU = `false`
	)

Referenced by ClassImp().

◆ estimate_CPU()

void URANIE::Modeler::TKriging::estimate_CPU	(	URANIE::DataServer::TDataServer *	tdsQuery,
		TString	listOfInputs = `""`,
		TString	listOfOutputs = `""`,
		Option_t *	option = `""`
	)

Referenced by ClassImp().

◆ estimateWithCov()

void URANIE::Modeler::TKriging::estimateWithCov	(	URANIE::DataServer::TDataServer *	tdsQuery,
		TString	listOfInputs = `""`,
		TString	listOfOutputs = `""`,
		TString	realValue = `""`,
		Option_t *	option = `""`
	)

Referenced by ClassImp().

◆ eval()

Int_t URANIE::Modeler::TKriging::eval	(	Double_t *	x0,
		Double_t *	y0,
		int	= `0`
	)

Evaluate the gaussian process at the new location x0.

This function aims to be compatible with the TEval prototype, which will allow to call it from a launcher or an optimizer.

Parameters

x0	(Double_t*): a single set of input variables values.
y0	(Double_t*): output of the gaussian process for x0.

Warning: the size of x0 is not checked. If it is different from the expected number of input variables, this will lead to unexpected results or an error.

Note: Calls the gpPred0 function from the gpL

Referenced by ClassImp().

◆ exportFunction()

void URANIE::Modeler::TKriging::exportFunction	(	const char *	lang,
		const char *	file,
		const char *	name,
		Option_t *	option = `""`
	)

Referenced by ClassImp().

◆ exportModelCplusplus()

virtual void URANIE::Modeler::TKriging::exportModelCplusplus	(	std::ofstream *	sourcefile,
		const char *	name = `""`
	)

virtual

Export the model as a C function (not implemented)

Referenced by ClassImp().

◆ exportModelFortran()

virtual void URANIE::Modeler::TKriging::exportModelFortran ( std::ofstream * sourcefile )

inlinevirtual

Export the model as a Fortran function (not implemented)

Referenced by ClassImp().

◆ exportModelPMML()

virtual void URANIE::Modeler::TKriging::exportModelPMML	(	const char *	file,
		const char *	name,
		Option_t *	option
	)		const

virtual

Referenced by ClassImp().

◆ freeze()

void URANIE::Modeler::TKriging::freeze ( )

Referenced by ClassImp().

◆ getCorrFunction()

TCorrelationFunction * URANIE::Modeler::TKriging::getCorrFunction ( )

inline

Return a pointer to the correlation function of the gaussian process.

References _correlationFunction.

◆ getCorrFunctionParams()

Double_t * URANIE::Modeler::TKriging::getCorrFunctionParams ( )

inline

Return all the parameters of the correlation function.

Warning: this is an alias to TCorrelationFunction::getParameters(). This means that the correlation lengths are stored at the end of the array, and are normalised.

References _correlationFunction, and URANIE::Modeler::TCorrelationFunction::getParameters().

Referenced by ClassImp().

◆ getCorrLengths()

Double_t * URANIE::Modeler::TKriging::getCorrLengths ( )

inline

Return the correlation lengths.

Note: these are NOT the correlation lengths stored in the correlation function. These values are denormalised in order to be interpretable in the "real" input space.

References _corrLengths.

Referenced by ClassImp().

◆ getCorrLengthsNormalised()

Double_t * URANIE::Modeler::TKriging::getCorrLengthsNormalised ( )

inline

Return the normalised correlation lengths.

Note: this corresponds to the correlation lengths stored in the correlation function.

References _normCorrLengths.

Referenced by ClassImp().

◆ getGPParams()

Double_t * URANIE::Modeler::TKriging::getGPParams ( )

inline

Return the array of the Gaussian Process parameters.

References _gamma.

◆ getiL()

Double_t * URANIE::Modeler::TKriging::getiL ( )

inline

Return the iL matrix as an array (cf. internal variable description)

References _iL.

◆ getInputNames()

const char * URANIE::Modeler::TKriging::getInputNames ( )

inline

Return the input variable names.

References _inputVarNames.

◆ getiR()

Double_t * URANIE::Modeler::TKriging::getiR ( )

inline

Return the iR matrix as an array (cf. internal variable description)

References _iR.

◆ getLooData()

URANIE::DataServer::TDataServer * URANIE::Modeler::TKriging::getLooData ( )

Return a pointer to a TDataServer containing the observations and the Leave One Out results.

Note: the user is responsible of the memory allocated for the data server.

Referenced by ClassImp().

◆ getLooErrors() [1/2]

Double_t * URANIE::Modeler::TKriging::getLooErrors ( )

inline

Return the vector of the leave one out errors.

References _errorLoo.

◆ getLooErrors() [2/2]

void URANIE::Modeler::TKriging::getLooErrors	(	double *	arr,
		int	size
	)

inline

Return the vector of the leave one out errors.

References _errorLoo, and _nS.

◆ getLooQ2()

Double_t URANIE::Modeler::TKriging::getLooQ2 ( )

inline

Return the Q2 of the model on the observation.

References _mseLoo, _nS, and _yObs.

Referenced by ClassImp().

◆ getLooRMSE()

Double_t URANIE::Modeler::TKriging::getLooRMSE ( )

inline

Return the leave one out mean squared error.

References _mseLoo.

Referenced by ClassImp().

◆ getLooVariances() [1/2]

Double_t * URANIE::Modeler::TKriging::getLooVariances ( )

inline

Return the vector of the leave one out prediction variances.

References _varLoo.

◆ getLooVariances() [2/2]

void URANIE::Modeler::TKriging::getLooVariances	(	double *	arr,
		int	size
	)

inline

Return the vector of the leave one out prediction variances.

References _nS, and _varLoo.

◆ getM()

Double_t * URANIE::Modeler::TKriging::getM ( )

inline

Return the M matrix as an array (cf. internal variable description)

References _M.

◆ getMeasurementErrorVariance()

Double_t URANIE::Modeler::TKriging::getMeasurementErrorVariance ( )

inline

Return the variance of the measurement error.

Note: if a value of -1.0 is returned, it means that the measurement error used to create the model was user defined. It was not considered useful to saved it with the model.

References _vErrMes.

◆ getNCorrFunctionParams()

Int_t URANIE::Modeler::TKriging::getNCorrFunctionParams ( )

inline

Return the number of parameters of the correlation function.

References _correlationFunction, and URANIE::Modeler::TCorrelationFunction::getNParameters().

Referenced by ClassImp().

◆ getNCorrLengths()

Int_t URANIE::Modeler::TKriging::getNCorrLengths ( )

inline

Return the number of correlation lengths.

References _correlationFunction, and URANIE::Modeler::TCorrelationFunction::getNCorrLengths().

Referenced by ClassImp().

◆ getNInputs()

Int_t URANIE::Modeler::TKriging::getNInputs ( )

inline

Return the number of input parameters.

References _nX.

◆ getNObs()

Int_t URANIE::Modeler::TKriging::getNObs ( )

inline

Return the number of observations used to build the GP model.

References _nS.

◆ getNTrendParams()

Int_t URANIE::Modeler::TKriging::getNTrendParams ( )

inline

Return the number of parameters of the deterministic trend.

References _nP.

◆ getObsInputs()

Double_t * URANIE::Modeler::TKriging::getObsInputs ( )

inline

Return the observations inputs as a flat array: [a1, b1, c1, a2, b2, c2,...].

References _xObs.

◆ getObsNormParams()

Double_t * URANIE::Modeler::TKriging::getObsNormParams ( )

inline

Return the normalisation parameters of the inputs as a flat array: [aMin, aMax, bMin, bMax,...].

References _xNormParams.

◆ getObsOutputs()

Double_t * URANIE::Modeler::TKriging::getObsOutputs ( )

inline

Return the observations outputs.

References _yObs.

◆ getOutputName()

const char * URANIE::Modeler::TKriging::getOutputName ( )

inline

Return the output variable name.

References _outputVarName.

◆ getTrendFormula()

TObjArray * URANIE::Modeler::TKriging::getTrendFormula ( )

inline

Return the deterministic trend list of formulas.

References _trendCoefList.

◆ getTrendParams()

Double_t * URANIE::Modeler::TKriging::getTrendParams ( )

inline

Return the array of the trend parameters.

References _hbeta.

◆ getTrendParamsCovMatrix()

Double_t * URANIE::Modeler::TKriging::getTrendParamsCovMatrix ( )

inline

Return the array of the trend parameters covariance matrix.

References _vbeta.

◆ getTrendString()

TString URANIE::Modeler::TKriging::getTrendString ( )

inline

Return the deterministic trend string.

References _trendString.

◆ getVariance()

Double_t URANIE::Modeler::TKriging::getVariance ( )

inline

Return the variance of the gaussian process.

References _sigma2.

◆ hasTrend()

Bool_t URANIE::Modeler::TKriging::hasTrend ( )

inline

Return kTRUE if a deterministic trend has been defined.

References bHas_trend.

◆ initMatrices()

void URANIE::Modeler::TKriging::initMatrices ( )

Initialisation of the matrices.

Referenced by ClassImp().

◆ printLog()

virtual void URANIE::Modeler::TKriging::printLog ( Option_t * option = "" )

virtual

Referenced by ClassImp().

◆ setCorrFunction()

void URANIE::Modeler::TKriging::setCorrFunction ( TCorrelationFunction * corrFunc )

inline

Set the correlation function of the gaussian process.

Parameters

corrFunc (URANIE::Modeler::TCorrelationFunction*): a pointer to the correlation function of the gaussian process.

References _correlationFunction.

◆ setLooErrors()

void URANIE::Modeler::TKriging::setLooErrors	(	Double_t *	err,
		Double_t *	var,
		Double_t	mse
	)

Set the information about the leave one out error of the model.

This function is called by the GPBuilder if the user wants the leave one out error of the model. The errors are computed thanks to the gpLooXError functions of the gpLib.

Parameters

err	(Double_t*): vector of size _nS of the errors from Leave One Out.
var	(Double_t*): vector of size _nS of the analytical variances of the leave one out errors.
mse	(Double_t): leave one out mean squared error.

Referenced by ClassImp(), and ClassImp().

◆ setVariance()

void URANIE::Modeler::TKriging::setVariance ( Double_t var )

inline

Set the variance of the gaussian process.

Parameters

var	(Double_t): value of the variance.

References _sigma2.

◆ StatMat() [1/2]

static void URANIE::Modeler::TKriging::StatMat	(	std::string	title,
		int	n1,
		int	n2,
		Double_t *	v
	)

static

Referenced by ClassImp().

◆ StatMat() [2/2]

static void URANIE::Modeler::TKriging::StatMat	(	std::string	title,
		int	n1,
		int	n2,
		Double_t *	v,
		std::vector< double > &	tvmin,
		std::vector< double > &	tvmax,
		std::vector< double > &	tmean,
		std::vector< double > &	tvar,
		std::vector< double > &	tnorm
	)

static

◆ StatVect() [1/2]

static void URANIE::Modeler::TKriging::StatVect	(	std::string	title,
		int	n,
		Double_t *	v
	)

static

Referenced by ClassImp().

◆ StatVect() [2/2]

static void URANIE::Modeler::TKriging::StatVect	(	std::string	title,
		int	n,
		Double_t *	v,
		double &	vmin,
		double &	vmax,
		double &	mean,
		double &	var,
		double &	norm
	)

static

Member Data Documentation

◆ _blog

Bool_t URANIE::Modeler::TKriging::_blog

private

Boolean for edit the log.

Referenced by ClassImp().

◆ _correlationFunction

TCorrelationFunction* URANIE::Modeler::TKriging::_correlationFunction

The correlation function.

Referenced by ClassImp(), getCorrFunction(), getCorrFunctionParams(), getNCorrFunctionParams(), getNCorrLengths(), and setCorrFunction().

◆ _corrLengths

Double_t* URANIE::Modeler::TKriging::_corrLengths

The correlation lengths in the real input space.

Referenced by ClassImp(), and getCorrLengths().

◆ _errorLoo

Double_t* URANIE::Modeler::TKriging::_errorLoo

If available, vector of the Leave One Out errors for the model.

Referenced by ClassImp(), getLooErrors(), and getLooErrors().

◆ _gamma

Double_t* URANIE::Modeler::TKriging::_gamma

Array of parameters of the gaussian process (size: _nS)

Referenced by ClassImp(), and getGPParams().

◆ _hbeta

Double_t* URANIE::Modeler::TKriging::_hbeta

Estimates of the deterministic trend parameters (size: _nP)

Referenced by ClassImp(), and getTrendParams().

◆ _iL

Double_t* URANIE::Modeler::TKriging::_iL

Inverse of the matrix L, Cholesky decomposition of K (size: _nS * _nS)

Referenced by ClassImp(), and getiL().

◆ _inputNamesList

TObjArray* URANIE::Modeler::TKriging::_inputNamesList

list of the inputs' names.

Referenced by ClassImp().

◆ _inputVarNames

TString URANIE::Modeler::TKriging::_inputVarNames

Name of the inputs variables separated by ":". This information can be used to remind the user of the original names of the variables.

Referenced by ClassImp(), and getInputNames().

◆ _iR

Double_t* URANIE::Modeler::TKriging::_iR

Inverse of the matrix R, from the QR decomposition of M (size: _nP * _nP)

Referenced by ClassImp(), and getiR().

◆ _M

Double_t* URANIE::Modeler::TKriging::_M

Matrix M = L^{-T} H (size: _nS * _nP)

Referenced by ClassImp(), and getM().

◆ _mseLoo

Double_t URANIE::Modeler::TKriging::_mseLoo

if available, Leave one Out mean squared error of the model.

Referenced by ClassImp(), getLooQ2(), and getLooRMSE().

◆ _normCorrLengths

Double_t* URANIE::Modeler::TKriging::_normCorrLengths

The correlation lengths in the normalised input space. This corresponds to the correlation lengths stored in the correlation function.

Referenced by ClassImp(), and getCorrLengthsNormalised().

◆ _nP

Int_t URANIE::Modeler::TKriging::_nP

Number of coefficients of the deterministic trend.

Referenced by ClassImp(), and getNTrendParams().

◆ _nS

Int_t URANIE::Modeler::TKriging::_nS

Number of observations.

Referenced by ClassImp(), getLooErrors(), getLooQ2(), getLooVariances(), and getNObs().

◆ _nX

Int_t URANIE::Modeler::TKriging::_nX

Number of input attributes

Referenced by ClassImp(), and getNInputs().

◆ _nXDeclared

Int_t URANIE::Modeler::TKriging::_nXDeclared

Number of attribute added to the tkiging object before running it.

Referenced by ClassImp().

◆ _nYDeclared

Int_t URANIE::Modeler::TKriging::_nYDeclared

Number of attribute added to the tkiging object before running it.

Referenced by ClassImp().

◆ _outputVarName

TString URANIE::Modeler::TKriging::_outputVarName

Name of the output variable. This information can be used to remind the user of the original names of the variables.

Referenced by ClassImp(), and getOutputName().

◆ _sigma2

Double_t URANIE::Modeler::TKriging::_sigma2

Variance of the gaussian process.

Referenced by ClassImp(), getVariance(), and setVariance().

◆ _trendCoefList

TObjArray* URANIE::Modeler::TKriging::_trendCoefList

List of the formulas of the trend coefficients (_nP elements)

Referenced by ClassImp(), and getTrendFormula().

◆ _trendString

TString URANIE::Modeler::TKriging::_trendString

Character string containing the trend parameters separated by ":" or a trend type ("const" or "linear")

Referenced by ClassImp(), and getTrendString().

◆ _varLoo

Double_t* URANIE::Modeler::TKriging::_varLoo

If available, vector of the variances of the Leave One Out errors for the model.

Referenced by ClassImp(), getLooVariances(), and getLooVariances().

◆ _vbeta

Double_t* URANIE::Modeler::TKriging::_vbeta

Covariance matrix of the trend parameters (size: _nP*_nP)

Referenced by ClassImp(), and getTrendParamsCovMatrix().

◆ _vErrMes

Double_t URANIE::Modeler::TKriging::_vErrMes

variance of the measurement error

Referenced by ClassImp(), and getMeasurementErrorVariance().

◆ _xNormParams

Double_t* URANIE::Modeler::TKriging::_xNormParams

The matrix of normalisation parameters. Contains min and max value of each input variable: [aMin, aMax, bMin, bMax,...] (size: _nX * 2)

Referenced by ClassImp(), and getObsNormParams().

◆ _xObs

Double_t* URANIE::Modeler::TKriging::_xObs

The matrix of inputs (dimension: _nX * _nS)

Referenced by ClassImp(), and getObsInputs().

◆ _yObs

Double_t* URANIE::Modeler::TKriging::_yObs

Array representation of the outputs (size: _nS)

Referenced by ClassImp(), getLooQ2(), and getObsOutputs().

◆ bHas_trend

Bool_t URANIE::Modeler::TKriging::bHas_trend

kFALSE if no deterministic trend is defined, kTRUE otherwise

Referenced by ClassImp(), and hasTrend().

◆ bUse_normalisation

Bool_t URANIE::Modeler::TKriging::bUse_normalisation

if kFALSE, the matrix H will be computed with unormalised input data. This happens when the user is providing his own trend function. When the trend is constructed automatically from a trend type, this attribute is set to kTRUE.

Referenced by ClassImp().

Documentation / Developer's manual

Available modules

Public Member Functions

Public Attributes

Private Attributes

Set methods

Detailed Description

Constructor & Destructor Documentation

◆ TKriging()

◆ ~TKriging()

Member Function Documentation

◆ computeCovarianceMatrix()

◆ createTrend()

◆ estimate()

◆ estimate_CPU()

◆ estimateWithCov()

◆ eval()

◆ exportFunction()

◆ exportModelCplusplus()

◆ exportModelFortran()

◆ exportModelPMML()

◆ freeze()

◆ getCorrFunction()

◆ getCorrFunctionParams()

◆ getCorrLengths()

◆ getCorrLengthsNormalised()

◆ getGPParams()

◆ getiL()

◆ getInputNames()

◆ getiR()

◆ getLooData()

◆ getLooErrors() [1/2]

◆ getLooErrors() [2/2]

◆ getLooQ2()

◆ getLooRMSE()

◆ getLooVariances() [1/2]

◆ getLooVariances() [2/2]

◆ getM()

◆ getMeasurementErrorVariance()

◆ getNCorrFunctionParams()

◆ getNCorrLengths()

◆ getNInputs()

◆ getNObs()

◆ getNTrendParams()

◆ getObsInputs()

◆ getObsNormParams()

◆ getObsOutputs()

◆ getOutputName()

◆ getTrendFormula()

◆ getTrendParams()

◆ getTrendParamsCovMatrix()

◆ getTrendString()

◆ getVariance()

◆ hasTrend()

◆ initMatrices()

◆ printLog()

◆ setCorrFunction()

◆ setLooErrors()

◆ setVariance()

◆ StatMat() [1/2]

◆ StatMat() [2/2]

◆ StatVect() [1/2]

◆ StatVect() [2/2]

Member Data Documentation

◆ _blog

◆ _correlationFunction

◆ _corrLengths

◆ _errorLoo

◆ _gamma

◆ _hbeta

◆ _iL

◆ _inputNamesList

◆ _inputVarNames

◆ _iR

◆ _M

◆ _mseLoo

◆ _normCorrLengths

◆ _nP

◆ _nS

◆ _nX