The abstract interface for the objects that control the calculation of the likelihood. More...

#include <LauFitObject.hh>

Inheritance diagram for LauFitObject:

Classes
struct	FormulaConstraint
	Struct to store formula-based constraint information. More...

class	MultiDimConstraint
	Class to store n-dimensional constraint information. More...

Public Member Functions
virtual	~LauFitObject ()=default
	Destructor.

void	useAsymmFitErrors (Bool_t useAsymmErrors)
	Turn on or off the computation of asymmetric errors (e.g. MINOS routine in Minuit) More...

Bool_t	useAsymmFitErrors () const
	Report whether or not calculation of asymmetric errors is enabled.

void	twoStageFit (Bool_t doTwoStageFit)
	Turn on or off the two stage fit. More...

Bool_t	twoStageFit () const
	Report whether the two-stage fit is enabled.

virtual void	withinAsymErrorCalc (const Bool_t inAsymErrCalc)
	Mark that the fit is calculating asymmetric errors. More...

Bool_t	withinAsymErrorCalc () const
	Query whether the fit is calculating the asymmetric errors. More...

void	setNExpts (UInt_t nExperiments, UInt_t firstExperiment, Bool_t toyExpts)
	Set the number of experiments, the first experiment, and whether this is toy. More...

UInt_t	eventsPerExpt () const
	Obtain the total number of events in the current experiment.

UInt_t	nExpt () const
	Obtain the number of experiments.

UInt_t	firstExpt () const
	Obtain the number of the first experiment.

UInt_t	iExpt () const
	Obtain the number of the current experiment.

Bool_t	toyExpts () const
	Obtain whether this is toy.

virtual void	setParsFromMinuit (Double_t *par, Int_t npar)=0
	This function sets the parameter values from Minuit. More...

virtual Double_t	getTotNegLogLikelihood ()=0
	Calculate the new value of the negative log likelihood. More...

void	addConstraint (const TString &formula, const std::vector< TString > &pars, const Double_t mean, const Double_t width)
	Store constraint information for fit parameters. More...

void	addFormulaConstraint (const TString &formula, const std::vector< TString > &pars, const Double_t mean, const Double_t width)
	Store constraint information for fit parameters. More...

void	addMultiDimConstraint (const std::vector< TString > &pars, const TVectorD &means, const TMatrixD &covMat)
	Store n-dimensional constraint information for fit parameters. More...

Protected Types
enum	ConstraintType { ConstraintType::Formula, ConstraintType::MultDim }
	Enumeration of the different types of constraint. More...

Protected Member Functions
	LauFitObject ()
	Constructor.

Bool_t	checkRepetition (const std::vector< TString > &names, const ConstraintType conType)
	Check if parameters names for constraints have already been used elsewhere. More...

void	generateConstraintMeans (std::vector< LauAbsRValue * > &conVars)
	Generate per-experiment mean for each Gaussian constraint. More...

const std::vector< FormulaConstraint > &	formulaConstraints () const
	Const access to the formula constraints store.

std::vector< FormulaConstraint > &	formulaConstraints ()
	Access to the formula constraints store.

const std::vector< MultiDimConstraint > &	multiDimConstraints () const
	Const access to the ND constraints store.

std::vector< MultiDimConstraint > &	multiDimConstraints ()
	Access to the ND constraints store.

const std::set< TString > &	formulaConstrainedPars () const
	Const access to the parameter names used in formula constraints.

std::set< TString > &	formulaConstrainedPars ()
	Access to the parameter names used in formula constraints.

const std::set< TString > &	multiDimConstrainedPars () const
	Const access to the parameter names used in ND constraints.

std::set< TString > &	multiDimConstrainedPars ()
	Access to the parameter names used in ND constraints.

void	resetFitCounters ()
	Reset the good/bad fit counters.

void	setCurrentExperiment (const UInt_t curExpt)
	Set the ID of the current experiment. More...

void	startNewFit (const UInt_t nPars, const UInt_t nFreePars)
	Indicate the start of a new fit. More...

void	eventsPerExpt (UInt_t nEvents)
	Set the number of events in the current experiment.

Double_t	worstLogLike () const
	Access the worst log likelihood found so far.

void	worstLogLike (const Double_t newWorstLogLike)
	Set a new value for the worst log likelihood. More...

void	storeFitStatus (const LauAbsFitter::FitStatus &status, const TMatrixD &covMatrix)
	Store fit status information. More...

UInt_t	nTotParams () const
	Access the total number of fit parameters.

UInt_t	nFreeParams () const
	Access the total number of fit parameters.

const LauAbsFitter::FitStatus &	fitStatus () const
	Access the fit status information.

Double_t	nll () const
	Access the current NLL value.

Double_t	edm () const
	Access the current EDM value.

Int_t	statusCode () const
	Access the fit status code.

const TMatrixD &	covarianceMatrix () const
	Access the fit covariance matrix.

UInt_t	numberOKFits () const
	Access the number of successful fits.

UInt_t	numberBadFits () const
	Access the number of failed fits.

Private Member Functions
	LauFitObject (const LauFitObject &rhs)
	Copy constructor (not implemented)

LauFitObject &	operator= (const LauFitObject &rhs)
	Copy assignment operator (not implemented)

Private Attributes
std::vector< FormulaConstraint >	formulaConstraints_
	Store the constraints for fit parameters until initialisation is complete.

std::vector< MultiDimConstraint >	multiDimConstraints_
	Store the ND constraints for fit parameters until initialisation is complete.

std::set< TString >	formulaConstrainedPars_
	Store the names of all parameters used in all formula constraints.

std::set< TString >	multiDimConstrainedPars_
	Store the names of all parameters used in all multi-dimensional constraints.

Bool_t	twoStageFit_
	Option to perform a two stage fit.

Bool_t	useAsymmFitErrors_
	Option to use asymmetric errors.

UInt_t	nParams_
	The number of fit parameters.

UInt_t	nFreeParams_
	The number of free fit parameters.

Bool_t	withinAsymErrorCalc_
	Flag to indicate if the asymmetric error calculation (e.g. MINOS) is currently running.

Bool_t	toyExpts_
	Flag to indicate whether this is toy. More...

UInt_t	firstExpt_
	The number of the first experiment to consider.

UInt_t	nExpt_
	The number of experiments to consider.

UInt_t	iExpt_
	The current experiment number.

UInt_t	evtsPerExpt_
	The number of events in the current experiment.

LauAbsFitter::FitStatus	fitStatus_
	The status of the current fit.

Double_t	worstLogLike_
	The worst log likelihood value found so far.

TMatrixD	covMatrix_
	The fit covariance matrix.

UInt_t	numberOKFits_
	The number of successful fits.

UInt_t	numberBadFits_
	The number of fit failures.

Detailed Description

The abstract interface for the objects that control the calculation of the likelihood.

Definition at line 47 of file LauFitObject.hh.

Member Enumeration Documentation

◆ ConstraintType

enum LauFitObject::ConstraintType

strongprotected

Enumeration of the different types of constraint.

Enumerator
Formula	Formula-based constraint on a combination of parameters.
MultDim	Multi-dimensional constraint on several parameters.

Definition at line 184 of file LauFitObject.hh.

Member Function Documentation

◆ addConstraint()

void LauFitObject::addConstraint	(	const TString &	formula,
		const std::vector< TString > &	pars,
		const Double_t	mean,
		const Double_t	width
	)

Store constraint information for fit parameters.

Deprecated:: Renamed to addFormulaConstraint, please switch to use this. Will be dropped in next major release.

Parameters

[in]	formula	the formula to be used in the LauFormulaPar
[in]	pars	a vector of LauParameter names to be used in the Formula, in the order specified by the formula
[in]	mean	the value of the mean of the Gaussian constraint
[in]	width	the value of the width of the Gaussian constraint

Definition at line 114 of file LauFitObject.cc.

◆ addFormulaConstraint()

void LauFitObject::addFormulaConstraint	(	const TString &	formula,
		const std::vector< TString > &	pars,
		const Double_t	mean,
		const Double_t	width
	)

Store constraint information for fit parameters.

Parameters

[in]	formula	the formula to be used in the LauFormulaPar
[in]	pars	a vector of LauParameter names to be used in the Formula, in the order specified by the formula
[in]	mean	the value of the mean of the Gaussian constraint
[in]	width	the value of the width of the Gaussian constraint

Definition at line 124 of file LauFitObject.cc.

◆ addMultiDimConstraint()

void LauFitObject::addMultiDimConstraint	(	const std::vector< TString > &	pars,
		const TVectorD &	means,
		const TMatrixD &	covMat
	)

Store n-dimensional constraint information for fit parameters.

Parameters

[in]	pars	a vector of LauParameter names to be used in the constraint
[in]	means	the values of the means of the Gaussian constraint
[in]	covMat	the covariance matrix of the parameters of the Gaussian constraint

Definition at line 139 of file LauFitObject.cc.

◆ checkRepetition()

Bool_t LauFitObject::checkRepetition	(	const std::vector< TString > &	names,
		const ConstraintType	conType
	)

protected

Check if parameters names for constraints have already been used elsewhere.

Also adds the names to the appropriate set

Parameters

[in]	names	a vector of parameter names
[in]	conType	the type of constraint being added - determines which set to which to add the parmaeters

Returns: kTRUE if no repetitions found, kFALSE if one or more repetitions found

Definition at line 177 of file LauFitObject.cc.

◆ generateConstraintMeans()

void LauFitObject::generateConstraintMeans ( std::vector< LauAbsRValue * > & conVars )

protected

Generate per-experiment mean for each Gaussian constraint.

Generates a new mean for all Gaussian constraints. The constraints on a single fit parameter and the formula-based constraints are provided as the argument, while the multi-dimensional constraints are already a member variable.

Parameters

[in,out] conVars the fit parameter and formula-based constraints

Definition at line 154 of file LauFitObject.cc.

◆ getTotNegLogLikelihood()

virtual Double_t LauFitObject::getTotNegLogLikelihood ( )

pure virtual

Calculate the new value of the negative log likelihood.

This function has to be public since it is called from the global FCN. It should not be called otherwise!

Implemented in LauSimFitTask, LauASqMaxFinder, LauAbsFitModel, LauSimFitCoordinator, and LauRooFitTask.

◆ setCurrentExperiment()

void LauFitObject::setCurrentExperiment ( const UInt_t curExpt )

inlineprotected

Set the ID of the current experiment.

Parameters

[in] curExpt the experiment number

Definition at line 298 of file LauFitObject.hh.

◆ setNExpts()

void LauFitObject::setNExpts	(	UInt_t	nExperiments,
		UInt_t	firstExperiment,
		Bool_t	toyExpts
	)

Set the number of experiments, the first experiment, and whether this is toy.

The default settings are nExperiments = 1, firstExperiment = 0, toyExpts = kFALSE, i.e. the settings for fitting a single data sample. As such, this function only needs to be called if generating/fitting toy samples.

Parameters

[in]	nExperiments	the number of experiments
[in]	firstExperiment	the number of the first experiment
[in]	toyExpts	whether this is toy - determines whether to generate per-experiment means for each Gaussian constraint, as per arXiv:1210.7141

Definition at line 61 of file LauFitObject.cc.

◆ setParsFromMinuit()

virtual void LauFitObject::setParsFromMinuit	(	Double_t *	par,
		Int_t	npar
	)

pure virtual

This function sets the parameter values from Minuit.

This function has to be public since it is called from the global FCN. It should not be called otherwise!

Parameters

[in]	par	an array storing the various parameter values
[in]	npar	the number of free parameters

Implemented in LauSimFitTask, LauASqMaxFinder, LauAbsFitModel, LauSimFitCoordinator, and LauRooFitTask.

◆ startNewFit()

void LauFitObject::startNewFit	(	const UInt_t	nPars,
		const UInt_t	nFreePars
	)

protected

Indicate the start of a new fit.

Parameters

[in]	nPars	the total number of fit parameters
[in]	nFreePars	the number of free fit parameters

Definition at line 83 of file LauFitObject.cc.

◆ storeFitStatus()

void LauFitObject::storeFitStatus	(	const LauAbsFitter::FitStatus &	status,
		const TMatrixD &	covMatrix
	)

protected

Store fit status information.

Parameters

[in]	status	the status information of the fit
[in]	covMatrix	the fit covariance matrix

Definition at line 93 of file LauFitObject.cc.

◆ twoStageFit()

void LauFitObject::twoStageFit ( Bool_t doTwoStageFit )

inline

Turn on or off the two stage fit.

The two-stage fit allows certain parameters to be fixed in one stage and floated in another stage of the fit. Can be used, for example, in a CP fit where the CP-parameters are fixed to zero in the first stage (while the CP-average parameters are determined), then floated in the second.

Parameters

[in] doTwoStageFit boolean specifying whether or not the two-stage fit should be enabled

Definition at line 73 of file LauFitObject.hh.

◆ useAsymmFitErrors()

void LauFitObject::useAsymmFitErrors ( Bool_t useAsymmErrors )

inline

Turn on or off the computation of asymmetric errors (e.g. MINOS routine in Minuit)

Parameters

[in] useAsymmErrors boolean specifying whether or not the computation of asymmetric errors is enabled

Definition at line 57 of file LauFitObject.hh.

◆ withinAsymErrorCalc() [1/2]

Bool_t LauFitObject::withinAsymErrorCalc ( ) const

inline

Query whether the fit is calculating the asymmetric errors.

Returns: kTRUE if the fit is calculating the asymmetric errors, kFALSE otherwise

Definition at line 94 of file LauFitObject.hh.

◆ withinAsymErrorCalc() [2/2]

virtual void LauFitObject::withinAsymErrorCalc ( const Bool_t inAsymErrCalc )

inlinevirtual

Mark that the fit is calculating asymmetric errors.

This is called by the fitter interface to mark when entering and exiting the asymmetric error calculation.

Parameters

[in] inAsymErrCalc boolean marking that the fit is calculating the asymmetric errors

Reimplemented in LauSimFitCoordinator.

Definition at line 85 of file LauFitObject.hh.

◆ worstLogLike()

void LauFitObject::worstLogLike ( const Double_t newWorstLogLike )

inlineprotected

Set a new value for the worst log likelihood.

Parameters

[in] newWorstLogLike the new value of the worst log likelihood

Definition at line 317 of file LauFitObject.hh.

Member Data Documentation

◆ toyExpts_

Bool_t LauFitObject::toyExpts_

private

Flag to indicate whether this is toy.

Determines whether to generate per-experiment means for each Gaussian constraint, as per arXiv:1210.7141

Definition at line 389 of file LauFitObject.hh.

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

/home/epp/phsdba/Development/Laura++/master-dev/laura/inc/LauFitObject.hh
/home/epp/phsdba/Development/Laura++/master-dev/laura/src/LauFitObject.cc

Classes

Public Member Functions

Protected Types

Protected Member Functions

Private Member Functions

Private Attributes

Detailed Description

Member Enumeration Documentation

◆ ConstraintType

Member Function Documentation

◆ addConstraint()

◆ addFormulaConstraint()

◆ addMultiDimConstraint()

◆ checkRepetition()

◆ generateConstraintMeans()

◆ getTotNegLogLikelihood()

◆ setCurrentExperiment()

◆ setNExpts()

◆ setParsFromMinuit()

◆ startNewFit()

◆ storeFitStatus()

◆ twoStageFit()

◆ useAsymmFitErrors()

◆ withinAsymErrorCalc() [1/2]

◆ withinAsymErrorCalc() [2/2]

◆ worstLogLike()

Member Data Documentation

◆ toyExpts_