Description of the class TLinearBayesian. More...

#include <TLinearBayesian.h>

Inheritance diagram for URANIE::Calibration::TLinearBayesian:

Collaboration diagram for URANIE::Calibration::TLinearBayesian:

Public Member Functions
Constructor and Destructor
	TLinearBayesian (URANIE::DataServer::TDataServer tds, URANIE::Relauncher::TRun run, Int_t ns=1, Option_t *option="")
	Default constructor with TRun arg. More...

	TLinearBayesian (URANIE::DataServer::TDataServer tds, void(fcn)(Double_t , Double_t ), const char varexpinput, const char varexpoutput, int ns=1, Option_t *option="")
	Default Calibration constructor with the function argument: it contains the assessor to be used. More...

	TLinearBayesian (URANIE::DataServer::TDataServer tds, const char fcn, const char varexpinput, const char varexpoutput, int ns=1, Option_t *option="")
	Default Calibration constructor with the function argument: it contains the assessor to be used. More...

	TLinearBayesian (URANIE::DataServer::TDataServer tds, URANIE::Launcher::TCode fcode, int ns=1, const char *option="")
	Default Calibration constructor with the code argument: it contains the assessor to be used. More...

virtual	~TLinearBayesian ()
	Default destructor. More...

Predictions methods
void	computePredictionVariance (URANIE::DataServer::TDataServer *tds_new, string outname)
	Set parameter transformation function Sometimes the calibration is performed on transformed variables (for linear problem for instance) This method allows to provide a custom function to trasform the estimated parameters to the original one. the first argument of the fonction must an array of double of size _nPar and the second argument is also an array of double of size _nPar. More...

Setting and Getting attributs
void	setParameterTransformationFunction (void(fTransfoParam)(double , double *))
	Set parameter transformation function Sometimes the calibration is performed on transformed variables (for linear problem for instance) This method allows to provide a custom function to trasform the estimated parameters to the original one. the first argument of the fonction must an array of double of size _nPar and the second argument is also an array of double of size _nPar. More...

void	setRegressorName (const char *regName)
	Set the regressor matrix by providing the variables to be extracted from tdsObs. More...

void	setDistanceAndReference (const char funcName, URANIE::DataServer::TDataServer tdsRef, const char input, const char output, const char *weight="")
	Set the distance function and some needed informations. More...

void	setDistanceAndReference (URANIE::Calibration::TDistanceFunction distFunc, URANIE::DataServer::TDataServer tdsRef, const char input, const char output, const char *weight="")
	Set the distance function and some needed informations. More...

TMatrixD	getParameterValueMatrix ()
	get the matrix of parameter values More...

TMatrixD	getTransfParameterValueMatrix ()
	get the matrix of parameter values More...

TMatrixD	getParameterCovarianceMatrix ()
	get the matrix of parameter covariances More...

Visualisation methods
void	drawParameters (TString sTitre, const char variable="", const char select="1>0", Option_t option="")

void	printLog (Option_t *option="")
	dump content More...

Public Member Functions inherited from URANIE::Calibration::TCalibration
	TCalibration (URANIE::DataServer::TDataServer tds, URANIE::Relauncher::TRun run, int ns=1, Option_t *option="")
	Default constructor with the runner argument: it contains the assessor to be used. More...

	TCalibration (URANIE::DataServer::TDataServer tds, void(fcn)(Double_t , Double_t ), const char varexpinput, const char varexpoutput, int ns=1, Option_t *option="")
	Default Calibration constructor with the function argument: it contains the assessor to be used. More...

	TCalibration (URANIE::DataServer::TDataServer tds, const char fcn, const char varexpinput, const char varexpoutput, int ns=1, Option_t *option="")
	Default Calibration constructor with the function argument: it contains the assessor to be used. More...

	TCalibration (URANIE::DataServer::TDataServer tds, URANIE::Launcher::TCode fcode, int ns=1, const char *option="")
	Default Calibration constructor with the code argument: it contains the assessor to be used. More...

virtual	~TCalibration ()
	Default destructor. More...

void	estimateParameters (Option_t *option="")

void	estimateCustomResidues (string resName, int n_theta, double *theta)

void	checkReference (URANIE::DataServer::TDataServer tdsRef, const char input, const char output, const char weight)
	Check the consistency of the formation provided. More...

void	drawResidues (TString sTitre, const char variable="", const char select="1>0", Option_t option="")

void	setLog ()

void	unsetLog ()

void	changeLog ()

Bool_t	getLog ()

Int_t	getID ()

void	setSeed (UInt_t nval)
	Set the seed of the random generator if one is used. More...

UInt_t	getSeed ()
	Get the seed of the random generator if one is used. More...

const char *	getMethodName ()
	Get the method name. More...

void	setObservationCovarianceMatrix (TMatrixD &mat)
	Set the observatiton covariance matrix. More...

URANIE::Calibration::TDistanceFunction *	getDistanceFunction ()
	Return the distance function. More...

Int_t	getNPar ()
	Get the number of parameters to be calibrated. More...

URANIE::DataServer::TDataServer *	getEvaluationTDS ()
	Get the tds in which evaluation will be performed. More...

Protected Member Functions
void	computeParameters (Option_t *option="")
	internal method in which the estimation is performed for all inheriting classes More...

void	checktdsParContent ()

Protected Member Functions inherited from URANIE::Calibration::TCalibration
void	checkCanvasCreation (bool newcan)
	Create a canvas if needed. More...

void	initInputs ()
	Initialise some common inputs. More...

void	initResults (vector< string > *ParsedLines)
	Initialise some common inputs. More...

void	computeAPosterioriForDistribution ()
	Compute the a posteriori residual for many-solutions method. More...

virtual void	parseOption (Option_t *option="")
	Read the possible options. More...

void	setMethodName (const char *str)
	Set the Method name. More...

Protected Attributes
void(*	_fTransfoParam )(double , double )
	Parameter transformation function. More...

vector< double >	_aPrioriMode
	A priori modes of the laws. More...

TMatrixD	_mParCovariance
	Parameters covariance matrix. More...

TMatrixD	_mParValues
	Parametres matrix. More...

TMatrixD	_mTransfoParValues
	Parametres matrix. More...

TMatrixD	_mH
	Regressor matrix. More...

vector< string >	_vRegName
	Regressor names. More...

string	_regname
	regressor name More...

Protected Attributes inherited from URANIE::Calibration::TCalibration
URANIE::Calibration::TDistanceFunction *	_dFunc
	Pointer to chosen distance function. More...

URANIE::DataServer::TDSNtupleD *	_evTuple
	Pointer to the eval ntuple. More...

TList *	_listOfParameters
	List of the parameters to be calibrated. More...

TCanvas *	_canvas
	Canvas object to deal with. More...

TObjArray *	_drawingGarbageCollector
	Garbage collector for prints. More...

int	_nSam
	The number of sample in a posteriori distributions. More...

int	_nObs
	The number of observations in the reference database. More...

int	_nIterMax
	The maximum number of iteration allowed (meaning total number of code estimation is _nIterMax * _nObs) More...

int	_nPar
	Dimension of the parameters. More...

int	_nVar
	Dimension of the output and references to be compared with. More...

int	_nSeed
	The seed of the random generator. More...

TString	_sMethodName
	The method name. More...

TString	_referenceName
	The reference name. More...

TString	_outputName
	The output name. More...

vector< string >	_vrefName
	The reference names. More...

vector< string >	_voutName
	The output names. More...

TString	_weightName
	The weight name. More...

TMatrixD	_mObsCovMat
	Observation Covariance matrix. More...

bool	_buseMatrix
	Use matrix instead of vectors in the Distance Function. More...

bool	_bsaveAll
	Whether all evaluations should be saved, not only a priori and a posteriori. More...

bool	_bdontKeepAgreement
	Remove the agreement attribute from the tdsPar object. More...

bool	_buseMode
	Use Mode instead of Mean. More...

Bool_t	_blog
	Boolean for edit the log. More...

Additional Inherited Members
Public Types inherited from URANIE::Calibration::TCalibration
enum	ELauncher { kCode, kFunction, kRun, kUnknown }

Public Attributes inherited from URANIE::Calibration::TCalibration
URANIE::DataServer::TDataServer *	_tdsPar
	TDS containing parameters properties (parameters that should be calibrated) More...

URANIE::DataServer::TDataServer *	_tdsObs
	TDS containing observations used for calibration. More...

URANIE::DataServer::TDataServer *	_tdsEval
	TDS containing a priori / a posteriori evaluations. More...

ELauncher	_nLauncher
	The type of launcher. More...

TString	_sFunctionName
	The Name of the evaluatuor. More...

TString	_sEI
	The Name of input. More...

TString	_sEO
	The Name of output. More...

URANIE::Launcher::TCode *	_code
	The tcode. More...

URANIE::Relauncher::TRun *	_run
	Pointer to the runner to be used. More...

void(*	_pFunction )(double , double )
	Function pointer. More...

vector< URANIE::DataServer::TStochasticAttribute * >	_vatt
	internal vector of stochastic attribute for some methods More...

Detailed Description

Description of the class TLinearBayesian.

Constructor & Destructor Documentation

◆ TLinearBayesian() [1/4]

URANIE::Calibration::TLinearBayesian::TLinearBayesian	(	URANIE::DataServer::TDataServer *	tds,
		URANIE::Relauncher::TRun *	run,
		Int_t	ns = `1`,
		Option_t *	option = `""`
	)

Default constructor with TRun arg.

Default constructor with the runner argument: it contains the assessor to be used

Parameters

tds	: the dataserver that contains no data but only one attribute per parameter to be calibrated
run	: the runner that contains the assessor to be used on the reference data
ns	: number of sample to be generated (depending on the method)
option	: see parseOption below

Referenced by ClassImp().

◆ TLinearBayesian() [2/4]

URANIE::Calibration::TLinearBayesian::TLinearBayesian	(	URANIE::DataServer::TDataServer *	tds,
		void()(Double_t , Double_t *)	fcn,
		const char *	varexpinput,
		const char *	varexpoutput,
		int	ns = `1`,
		Option_t *	option = `""`
	)

Default Calibration constructor with the function argument: it contains the assessor to be used.

Parameters

tds	: the dataserver that contains no data but only one attribute per parameter to be calibrated
fcn	: the pointer to a function
ns	: number of sample to be generated (depending on the method)
varexpinput	: the input variable for the function in the correct order (both input and parameters)
varexpoutput	: the output of the function in the correct order

◆ TLinearBayesian() [3/4]

URANIE::Calibration::TLinearBayesian::TLinearBayesian	(	URANIE::DataServer::TDataServer *	tds,
		const char *	fcn,
		const char *	varexpinput,
		const char *	varexpoutput,
		int	ns = `1`,
		Option_t *	option = `""`
	)

Default Calibration constructor with the function argument: it contains the assessor to be used.

Parameters

tds	: the dataserver that contains no data but only one attribute per parameter to be calibrated
fcn	: the name of the function
ns	: number of sample to be generated (depending on the method)
varexpinput	: the input variable for the function in the correct order (both input and parameters)
varexpoutput	: the output of the function in the correct order

◆ TLinearBayesian() [4/4]

URANIE::Calibration::TLinearBayesian::TLinearBayesian	(	URANIE::DataServer::TDataServer *	tds,
		URANIE::Launcher::TCode *	fcode,
		int	ns = `1`,
		const char *	option = `""`
	)

Default Calibration constructor with the code argument: it contains the assessor to be used.

Parameters

tds	: the dataserver that contains no data but only one attribute per parameter to be calibrated
code	: the code object that will be runned
ns	: number of sample to be generated (depending on the method)

◆ ~TLinearBayesian()

virtual URANIE::Calibration::TLinearBayesian::~TLinearBayesian ( )

virtual

Default destructor.

Referenced by ClassImp().

Member Function Documentation

◆ checktdsParContent()

void URANIE::Calibration::TLinearBayesian::checktdsParContent ( )

protectedvirtual

Check the content of the input dataserver

Implements URANIE::Calibration::TCalibration.

Referenced by ClassImp().

◆ computeParameters()

void URANIE::Calibration::TLinearBayesian::computeParameters ( Option_t * option = "" )

protectedvirtual

internal method in which the estimation is performed for all inheriting classes

Implements URANIE::Calibration::TCalibration.

Referenced by ClassImp().

◆ computePredictionVariance()

void URANIE::Calibration::TLinearBayesian::computePredictionVariance	(	URANIE::DataServer::TDataServer *	tds_new,
		string	outname
	)

Set parameter transformation function Sometimes the calibration is performed on transformed variables (for linear problem for instance) This method allows to provide a custom function to trasform the estimated parameters to the original one. the first argument of the fonction must an array of double of size _nPar and the second argument is also an array of double of size _nPar.

Referenced by ClassImp().

◆ drawParameters()

void URANIE::Calibration::TLinearBayesian::drawParameters	(	TString	sTitre,
		const char *	variable = `"*"`,
		const char *	select = `"1>0"`,
		Option_t *	option = `""`
	)

virtual

brief Draws the parameters as functions (ROOT's TF1s) The estimateParameters method hass computed the gaussian parameters, so this one plots the parameters as continous function TF1

Parameters

sTitre	title of the plots if specified
variable	list of variables that should be plotted
select	useless for this method
option	possible option for this plot "nonewcanvas": don't create a new canvas, but use the one active "vertical" : if there are several parameters to be shown put them on top of another instead of side by side "transformed" : use the transformed central alues and not the nominal ones "same" : plot the function on the existing pad, using the existing axis.

Reimplemented from URANIE::Calibration::TCalibration.

Referenced by ClassImp().

◆ getParameterCovarianceMatrix()

TMatrixD URANIE::Calibration::TLinearBayesian::getParameterCovarianceMatrix ( )

inline

get the matrix of parameter covariances

References _mParCovariance.

◆ getParameterValueMatrix()

TMatrixD URANIE::Calibration::TLinearBayesian::getParameterValueMatrix ( )

inline

get the matrix of parameter values

References _mParValues.

◆ getTransfParameterValueMatrix()

TMatrixD URANIE::Calibration::TLinearBayesian::getTransfParameterValueMatrix ( )

inline

get the matrix of parameter values

References _mTransfoParValues.

◆ printLog()

void URANIE::Calibration::TLinearBayesian::printLog ( Option_t * option = "" )

virtual

dump content

Reimplemented from URANIE::Calibration::TCalibration.

Referenced by ClassImp().

◆ setDistanceAndReference() [1/2]

void URANIE::Calibration::TLinearBayesian::setDistanceAndReference	(	const char *	funcName,
		URANIE::DataServer::TDataServer *	tdsRef,
		const char *	input,
		const char *	output,
		const char *	weight = `""`
	)

virtual

Set the distance function and some needed informations.

Parameters

funcName	name of the distance function chosen among the already implemented ones: "LS" for least square "weighterLS" for weighted least square "relativeLS" for relative least square "L1" for L1 norm "Mahalanobis" for Mahalanobis distance function
tdsRef	: the dataserver that contains all needed information detailled below input: the input attributes used of the assessors in run reference: the reference attributes that will be used to compared the newly done estimations for every iterations weight: a weight to reweight every event, one-by-one (an experimental uncertainty for instance)

Reimplemented from URANIE::Calibration::TCalibration.

Referenced by ClassImp().

◆ setDistanceAndReference() [2/2]

void URANIE::Calibration::TLinearBayesian::setDistanceAndReference	(	URANIE::Calibration::TDistanceFunction *	distFunc,
		URANIE::DataServer::TDataServer *	tdsRef,
		const char *	input,
		const char *	output,
		const char *	weight = `""`
	)

virtual

Set the distance function and some needed informations.

Parameters

distFunc	a pointer to a TDistanceFunction object (not usual, recommended only when dealing with a home-made distance unction compiled on the spot)
tdsRef	: the dataserver that contains all needed information detailled below input: the input attributes used of the assessors in run reference: the reference attributes that will be used to compared the newly done estimations for every iterations weight: a weight to reweight every event, one-by-one (an experimental uncertainty for instance)

Reimplemented from URANIE::Calibration::TCalibration.

◆ setParameterTransformationFunction()

void URANIE::Calibration::TLinearBayesian::setParameterTransformationFunction ( void(*)(double *, double *) fTransfoParam )

inline

Set parameter transformation function Sometimes the calibration is performed on transformed variables (for linear problem for instance) This method allows to provide a custom function to trasform the estimated parameters to the original one. the first argument of the fonction must an array of double of size _nPar and the second argument is also an array of double of size _nPar.

References _fTransfoParam.

◆ setRegressorName()

void URANIE::Calibration::TLinearBayesian::setRegressorName ( const char * regName )

Set the regressor matrix by providing the variables to be extracted from tdsObs.

Referenced by ClassImp().

Member Data Documentation

◆ _aPrioriMode

vector<double> URANIE::Calibration::TLinearBayesian::_aPrioriMode

protected

A priori modes of the laws.

Referenced by ClassImp().

◆ _fTransfoParam

void(* URANIE::Calibration::TLinearBayesian::_fTransfoParam) (double *, double *)

protected

Parameter transformation function.

Referenced by ClassImp(), and setParameterTransformationFunction().

◆ _mH

TMatrixD URANIE::Calibration::TLinearBayesian::_mH

protected

Regressor matrix.

Referenced by ClassImp().

◆ _mParCovariance

TMatrixD URANIE::Calibration::TLinearBayesian::_mParCovariance

protected

Parameters covariance matrix.

Referenced by ClassImp(), and getParameterCovarianceMatrix().

◆ _mParValues

TMatrixD URANIE::Calibration::TLinearBayesian::_mParValues

protected

Parametres matrix.

Referenced by ClassImp(), and getParameterValueMatrix().

◆ _mTransfoParValues

TMatrixD URANIE::Calibration::TLinearBayesian::_mTransfoParValues

protected

Parametres matrix.

Referenced by ClassImp(), and getTransfParameterValueMatrix().

◆ _regname

string URANIE::Calibration::TLinearBayesian::_regname

protected

regressor name

Referenced by ClassImp().

◆ _vRegName

vector<string> URANIE::Calibration::TLinearBayesian::_vRegName

protected

Regressor names.

Referenced by ClassImp().

Documentation / Developer's manual

Available modules

Public Member Functions

Protected Member Functions

Protected Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ TLinearBayesian() [1/4]

◆ TLinearBayesian() [2/4]

◆ TLinearBayesian() [3/4]

◆ TLinearBayesian() [4/4]

◆ ~TLinearBayesian()

Member Function Documentation

◆ checktdsParContent()

◆ computeParameters()

◆ computePredictionVariance()

◆ drawParameters()

◆ getParameterCovarianceMatrix()

◆ getParameterValueMatrix()

◆ getTransfParameterValueMatrix()

◆ printLog()

◆ setDistanceAndReference() [1/2]

◆ setDistanceAndReference() [2/2]

◆ setParameterTransformationFunction()

◆ setRegressorName()

Member Data Documentation

◆ _aPrioriMode

◆ _fTransfoParam

◆ _mH

◆ _mParCovariance

◆ _mParValues

◆ _mTransfoParValues

◆ _regname

◆ _vRegName