MultiDimParamGradFunctionAdapter class to wrap a one-dimensional parametric gradient function in a multi dimensional parametric gradient function interface This is used typically in fitting where internally the function is stored as multidimensional.

To wrap a non-parametric one-dim gradient function in a multi-dim interface one can use simply a a ROOT::Math::GradFunctor

The parameters are not stored in the adapter class and by default the pointer to the 1D function is owned. This means that deleting the class deletes also the 1D function and copying the class copies also the 1D function This class differs from WrappedParamFunction in the fact that the parameters are not stored in the adapter class and optionally it keeps a cloned and managed copy of the adapter class.

Definition at line 172 of file MultiDimParamFunctionAdapter.h.

Public Types
typedef T	BackendType

typedef IParamMultiGradFunction::BaseFunc	BaseFunc

using	BaseGradFunc = IGradientFunctionMultiDimTempl<T>

using	BaseParamFunc = IParametricFunctionMultiDimTempl<T>

Public Member Functions
	MultiDimParamGradFunctionAdapter (const IParamGradFunction &f)
	Constructor from a param one dim function interface from a const reference Copy and manage the own function pointer.

	MultiDimParamGradFunctionAdapter (const MultiDimParamGradFunctionAdapter &rhs)
	Copy constructor.

	MultiDimParamGradFunctionAdapter (IParamGradFunction &f)
	Constructor from a param one dim function interface from a non const reference Do not own the function pointer in this case.

	~MultiDimParamGradFunctionAdapter () override
	Destructor (no operations)

BaseFunc *	Clone () const override
	clone

T	Derivative (const T x, unsigned int icoord, T previous_grad, T previous_g2, T previous_gstep) const
	In some cases, the derivative algorithm will use information from the previous step, these can be passed in with this overload.

T	Derivative (const T *x, unsigned int icoord=0) const
	Return the partial derivative with respect to the passed coordinate.

virtual void	FdF (const T x, T &f, T df) const
	Optimized method to evaluate at the same time the function value and derivative at a point x.

virtual void	Gradient (const T x, T grad) const
	Evaluate all the vector of function derivatives (gradient) at a point x.

virtual bool	HasGradient () const

virtual bool	HasParameterHessian () const

unsigned int	NDim () const override
	Retrieve the dimension of the function.

unsigned int	NPar () const override
	Return the number of Parameters.

T	operator() (const T *x) const

T	operator() (const T x, const double p) const

MultiDimParamGradFunctionAdapter &	operator= (const MultiDimParamGradFunctionAdapter &rhs)
	Assignment operator.

T	ParameterDerivative (const T x, const double p, unsigned int ipar=0) const
	Evaluate the partial derivative w.r.t a parameter ipar from values and parameters.

T	ParameterDerivative (const T *x, unsigned int ipar=0) const
	Evaluate partial derivative using cached parameter values.

virtual bool	ParameterG2 (const T , const double , T *) const
	Evaluate all the second derivatives (diagonal ones) of the function with respect to the parameters at a point x.

void	ParameterGradient (const double x, const double p, double *grad) const override

virtual void	ParameterGradient (const T x, const double p, T *grad) const
	Evaluate the all the derivatives (gradient vector) of the function with respect to the parameters at a point x.

void	ParameterGradient (const T x, T grad) const
	Evaluate all derivatives using cached parameter values.

virtual bool	ParameterHessian (const T , const double , T *) const
	Evaluate the all the Hessian (second derivatives matrix) of the function with respect to the parameters at a point x.

virtual std::string	ParameterName (unsigned int i) const
	Return the name of the i-th parameter (starting from zero) Overwrite if want to avoid the default name ("Par_0, Par_1, ...")

const double *	Parameters () const override
	Access the parameter values.

void	SetParameters (const double *p) override
	Set the parameter values.

Private Member Functions
virtual T	DoDerivative (const T *, unsigned int) const
	Function to evaluate the derivative with respect each coordinate. To be implemented by the derived class.

virtual T	DoDerivativeWithPrevResult (const T x, unsigned int icoord, T , T , T ) const
	In some cases, the derivative algorithm will use information from the previous step, these can be passed in with this overload.

T	DoEval (const T *x) const override
	Implement the ROOT::Math::IBaseFunctionMultiDim interface DoEval(x) using the cached parameter values.

double	DoEvalPar (const double x, const double p) const override
	functions needed by interface

double	DoParameterDerivative (const double x, const double p, unsigned int ipar) const override

Private Attributes
IParamGradFunction *	fFunc

bool	fOwn

#include <Math/MultiDimParamFunctionAdapter.h>

Inheritance diagram for ROOT::Math::MultiDimParamGradFunctionAdapter:

Member Typedef Documentation

◆ BackendType

template<class T >

typedef T ROOT::Math::IBaseFunctionMultiDimTempl< T >::BackendType

inherited

Definition at line 67 of file IFunction.h.

◆ BaseFunc

typedef IParamMultiGradFunction::BaseFunc ROOT::Math::MultiDimParamGradFunctionAdapter::BaseFunc

Definition at line 176 of file MultiDimParamFunctionAdapter.h.

◆ BaseGradFunc

template<class T >

using ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::BaseGradFunc = IGradientFunctionMultiDimTempl<T>

inherited

Definition at line 231 of file IParamFunction.h.

◆ BaseParamFunc

template<class T >

using ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::BaseParamFunc = IParametricFunctionMultiDimTempl<T>

inherited

Definition at line 230 of file IParamFunction.h.

Constructor & Destructor Documentation

◆ MultiDimParamGradFunctionAdapter() [1/3]

ROOT::Math::MultiDimParamGradFunctionAdapter::MultiDimParamGradFunctionAdapter ( const IParamGradFunction & f )

inline

Constructor from a param one dim function interface from a const reference Copy and manage the own function pointer.

Definition at line 183 of file MultiDimParamFunctionAdapter.h.

◆ MultiDimParamGradFunctionAdapter() [2/3]

ROOT::Math::MultiDimParamGradFunctionAdapter::MultiDimParamGradFunctionAdapter ( IParamGradFunction & f )

inline

Constructor from a param one dim function interface from a non const reference Do not own the function pointer in this case.

Definition at line 193 of file MultiDimParamFunctionAdapter.h.

◆ MultiDimParamGradFunctionAdapter() [3/3]

ROOT::Math::MultiDimParamGradFunctionAdapter::MultiDimParamGradFunctionAdapter ( const MultiDimParamGradFunctionAdapter & rhs )

inline

Copy constructor.

Different behaviour according if function is owned or not

Definition at line 202 of file MultiDimParamFunctionAdapter.h.

◆ ~MultiDimParamGradFunctionAdapter()

ROOT::Math::MultiDimParamGradFunctionAdapter::~MultiDimParamGradFunctionAdapter ( )

inlineoverride

Destructor (no operations)

Definition at line 215 of file MultiDimParamFunctionAdapter.h.

Member Function Documentation

◆ Clone()

BaseFunc * ROOT::Math::MultiDimParamGradFunctionAdapter::Clone ( ) const

inlineoverridevirtual

clone

Implements ROOT::Math::IBaseFunctionMultiDimTempl< T >.

Definition at line 239 of file MultiDimParamFunctionAdapter.h.

◆ Derivative() [1/2]

template<class T >

T ROOT::Math::IBaseFunctionMultiDimTempl< T >::Derivative	(	const T *	x,
		unsigned int	icoord,
		T *	previous_grad,
		T *	previous_g2,
		T *	previous_gstep ) const

inlineinherited

In some cases, the derivative algorithm will use information from the previous step, these can be passed in with this overload.

The previous_* arrays can also be used to return second derivative and step size so that these can be passed forward again as well at the call site, if necessary.

Definition at line 120 of file IFunction.h.

◆ Derivative() [2/2]

template<class T >

T ROOT::Math::IBaseFunctionMultiDimTempl< T >::Derivative	(	const T *	x,
		unsigned int	icoord = 0 ) const

inlineinherited

Return the partial derivative with respect to the passed coordinate.

Definition at line 115 of file IFunction.h.

◆ DoDerivative()

template<class T >

virtual T ROOT::Math::IBaseFunctionMultiDimTempl< T >::DoDerivative	(	const T *	,
		unsigned int	) const

inlineprivatevirtualinherited

Function to evaluate the derivative with respect each coordinate. To be implemented by the derived class.

Definition at line 131 of file IFunction.h.

◆ DoDerivativeWithPrevResult()

template<class T >

virtual T ROOT::Math::IBaseFunctionMultiDimTempl< T >::DoDerivativeWithPrevResult	(	const T *	x,
		unsigned int	icoord,
		T *	,
		T *	,
		T *	) const

inlineprivatevirtualinherited

In some cases, the derivative algorithm will use information from the previous step, these can be passed in with this overload.

The previous_* arrays can also be used to return second derivative and step size so that these can be passed forward again as well at the call site, if necessary.

Definition at line 136 of file IFunction.h.

◆ DoEval()

template<class T >

T ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::DoEval ( const T * x ) const

inlineoverrideprivatevirtualinherited

Implement the ROOT::Math::IBaseFunctionMultiDim interface DoEval(x) using the cached parameter values.

Reimplemented from ROOT::Math::IParametricFunctionMultiDimTempl< T >.

Reimplemented in ROOT::Math::WrappedMultiTF1Templ< T >.

Definition at line 307 of file IParamFunction.h.

◆ DoEvalPar()

double ROOT::Math::MultiDimParamGradFunctionAdapter::DoEvalPar	(	const double *	x,
		const double *	p ) const

inlineoverrideprivate

functions needed by interface

Definition at line 281 of file MultiDimParamFunctionAdapter.h.

◆ DoParameterDerivative()

double ROOT::Math::MultiDimParamGradFunctionAdapter::DoParameterDerivative	(	const double *	x,
		const double *	p,
		unsigned int	ipar ) const

inlineoverrideprivate

Definition at line 290 of file MultiDimParamFunctionAdapter.h.

◆ FdF()

template<class T >

virtual void ROOT::Math::IBaseFunctionMultiDimTempl< T >::FdF	(	const T *	x,
		T &	f,
		T *	df ) const

inlinevirtualinherited

Optimized method to evaluate at the same time the function value and derivative at a point x.

Often both value and derivatives are needed and it is often more efficient to compute them at the same time. Derived class should implement this method if performances play an important role and if it is faster to evaluate value and derivative at the same time

Definition at line 108 of file IFunction.h.

◆ Gradient()

template<class T >

virtual void ROOT::Math::IBaseFunctionMultiDimTempl< T >::Gradient	(	const T *	x,
		T *	grad ) const

inlinevirtualinherited

Evaluate all the vector of function derivatives (gradient) at a point x.

Derived classes must re-implement it if more efficient than evaluating one at a time

Definition at line 96 of file IFunction.h.

◆ HasGradient()

template<class T >

virtual bool ROOT::Math::IBaseFunctionMultiDimTempl< T >::HasGradient ( ) const

inlinevirtualinherited

Reimplemented in ROOT::Math::IGradientFunctionMultiDimTempl< T >.

Definition at line 92 of file IFunction.h.

◆ HasParameterHessian()

template<class T >

virtual bool ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::HasParameterHessian ( ) const

inlinevirtualinherited

Reimplemented in ROOT::Math::WrappedMultiTF1Templ< T >.

Definition at line 265 of file IParamFunction.h.

◆ NDim()

unsigned int ROOT::Math::MultiDimParamGradFunctionAdapter::NDim ( ) const

inlineoverridevirtual

Retrieve the dimension of the function.

Implements ROOT::Math::IBaseFunctionMultiDimTempl< T >.

Definition at line 262 of file MultiDimParamFunctionAdapter.h.

◆ NPar()

unsigned int ROOT::Math::MultiDimParamGradFunctionAdapter::NPar ( ) const

inlineoverridevirtual

Return the number of Parameters.

Implements ROOT::Math::IBaseParam.

Definition at line 257 of file MultiDimParamFunctionAdapter.h.

◆ operator()() [1/2]

template<class T >

T ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::operator() ( const T * x ) const

inlineinherited

Definition at line 248 of file IParamFunction.h.

◆ operator()() [2/2]

template<class T >

T ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::operator()	(	const T *	x,
		const double *	p ) const

inlineinherited

Definition at line 243 of file IParamFunction.h.

◆ operator=()

MultiDimParamGradFunctionAdapter & ROOT::Math::MultiDimParamGradFunctionAdapter::operator= ( const MultiDimParamGradFunctionAdapter & rhs )

inline

Assignment operator.

Definition at line 224 of file MultiDimParamFunctionAdapter.h.

◆ ParameterDerivative() [1/2]

template<class T >

T ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::ParameterDerivative	(	const T *	x,
		const double *	p,
		unsigned int	ipar = 0 ) const

inlineinherited

Evaluate the partial derivative w.r.t a parameter ipar from values and parameters.

Definition at line 283 of file IParamFunction.h.

◆ ParameterDerivative() [2/2]

template<class T >

T ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::ParameterDerivative	(	const T *	x,
		unsigned int	ipar = 0 ) const

inlineinherited

Evaluate partial derivative using cached parameter values.

Definition at line 295 of file IParamFunction.h.

◆ ParameterG2()

template<class T >

virtual bool ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::ParameterG2	(	const T *	,
		const double *	,
		T *	) const

inlinevirtualinherited

Evaluate all the second derivatives (diagonal ones) of the function with respect to the parameters at a point x.

g2 is a vector of dimension npar

Reimplemented in ROOT::Math::WrappedMultiTF1Templ< T >.

Definition at line 278 of file IParamFunction.h.

◆ ParameterGradient() [1/3]

void ROOT::Math::MultiDimParamGradFunctionAdapter::ParameterGradient	(	const double *	x,
		const double *	p,
		double *	grad ) const

inlineoverride

Definition at line 271 of file MultiDimParamFunctionAdapter.h.

◆ ParameterGradient() [2/3]

template<class T >

virtual void ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::ParameterGradient	(	const T *	x,
		const double *	p,
		T *	grad ) const

inlinevirtualinherited

Evaluate the all the derivatives (gradient vector) of the function with respect to the parameters at a point x.

It is optional to be implemented by the derived classes for better efficiency

Reimplemented in ROOT::Math::WrappedMultiTF1Templ< T >.

Definition at line 257 of file IParamFunction.h.

◆ ParameterGradient() [3/3]

template<class T >

void ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::ParameterGradient	(	const T *	x,
		T *	grad ) const

inlineinherited

Evaluate all derivatives using cached parameter values.

Definition at line 291 of file IParamFunction.h.

◆ ParameterHessian()

template<class T >

virtual bool ROOT::Math::IParametricGradFunctionMultiDimTempl< T >::ParameterHessian	(	const T *	,
		const double *	,
		T *	) const

inlinevirtualinherited

Evaluate the all the Hessian (second derivatives matrix) of the function with respect to the parameters at a point x.

It is optional to be implemented by the derived classes if needed. If it is not implemented return a false. h must be dimensioned as a n x (n+1)/2 matrix (since it is a symmetric matrix)

Reimplemented in ROOT::Math::WrappedMultiTF1Templ< T >.

Definition at line 272 of file IParamFunction.h.

◆ ParameterName()

virtual std::string ROOT::Math::IBaseParam::ParameterName ( unsigned int i ) const

inlinevirtualinherited

Return the name of the i-th parameter (starting from zero) Overwrite if want to avoid the default name ("Par_0, Par_1, ...")

Reimplemented in ROOT::Math::WrappedMultiTF1Templ< T >, ROOT::Math::WrappedTF1, ROOT::Math::VavilovAccurateCdf, ROOT::Math::VavilovAccuratePdf, and ROOT::Math::VavilovAccurateQuantile.

Definition at line 86 of file IParamFunction.h.

◆ Parameters()

const double * ROOT::Math::MultiDimParamGradFunctionAdapter::Parameters ( ) const

inlineoverridevirtual

Access the parameter values.

Implements ROOT::Math::IBaseParam.

Definition at line 247 of file MultiDimParamFunctionAdapter.h.

◆ SetParameters()

void ROOT::Math::MultiDimParamGradFunctionAdapter::SetParameters ( const double * p )

inlineoverridevirtual

Set the parameter values.

Parameters

p	vector of doubles containing the parameter values.

to be defined: can user change number of params ? At the moment no.

Implements ROOT::Math::IBaseParam.

Definition at line 252 of file MultiDimParamFunctionAdapter.h.

Member Data Documentation

◆ fFunc

IParamGradFunction* ROOT::Math::MultiDimParamGradFunctionAdapter::fFunc

private

Definition at line 298 of file MultiDimParamFunctionAdapter.h.

◆ fOwn

bool ROOT::Math::MultiDimParamGradFunctionAdapter::fOwn

private

Definition at line 297 of file MultiDimParamFunctionAdapter.h.

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

math/mathcore/inc/Math/MultiDimParamFunctionAdapter.h

Public Types

Public Member Functions

Private Member Functions

Private Attributes

Member Typedef Documentation

◆ BackendType

◆ BaseFunc

◆ BaseGradFunc

◆ BaseParamFunc

Constructor & Destructor Documentation

◆ MultiDimParamGradFunctionAdapter() [1/3]

◆ MultiDimParamGradFunctionAdapter() [2/3]

◆ MultiDimParamGradFunctionAdapter() [3/3]

◆ ~MultiDimParamGradFunctionAdapter()

Member Function Documentation

◆ Clone()

◆ Derivative() [1/2]

◆ Derivative() [2/2]

◆ DoDerivative()

◆ DoDerivativeWithPrevResult()

◆ DoEval()

◆ DoEvalPar()

◆ DoParameterDerivative()

◆ FdF()

◆ Gradient()

◆ HasGradient()

◆ HasParameterHessian()

◆ NDim()

◆ NPar()

◆ operator()() [1/2]

◆ operator()() [2/2]

◆ operator=()

◆ ParameterDerivative() [1/2]

◆ ParameterDerivative() [2/2]

◆ ParameterG2()

◆ ParameterGradient() [1/3]

◆ ParameterGradient() [2/3]

◆ ParameterGradient() [3/3]

◆ ParameterHessian()

◆ ParameterName()

◆ Parameters()

◆ SetParameters()

Member Data Documentation

◆ fFunc

◆ fOwn