The MultidimRegression class allows to regress a multidimensional output variable among a multivariate explanation variable. More...

#include <STK_MultidimRegression.h>

Inheritance diagram for STK::MultidimRegression< Array, Weight >:

Public Types
typedef IRegression< Array, Array, Weight >	Base

Public Member Functions
	MultidimRegression (Array const y=0, Array const x=0)
	Constructor.

virtual	~MultidimRegression ()
	Destructor.

Array const &	coefs () const

virtual Array	extrapolate (Array const &x) const

Public Member Functions inherited from STK::IRegression< Array, Array, Weight >
virtual	~IRegression ()
	virtual destructor.

Array const &	predicted () const

Array const &	residuals () const

Array *	p_predicted ()

Array *	p_residuals ()

int	nbFreeParameter () const

virtual bool	run ()
	run the computations.

virtual bool	run (Weight const &weights)
	run the weighted computations.

Public Member Functions inherited from STK::IRunnerSupervised< YArray_, XArray_, Weights_ >
virtual void	setX (XArray_ const &x)
	set the x data set (predictors).

virtual void	setY (YArray_ const &y)
	set the data set.

virtual void	setData (YArray_ const &y, XArray_ const &x)
	set the data set.

virtual bool	run (Weights_ const &weights)=0
	run the weighted computations.

Public Member Functions inherited from STK::IRunnerBase
String const &	error () const
	get the last error message.

Protected Attributes
ArrayXX	coefs_

Protected Attributes inherited from STK::IRegression< Array, Array, Weight >
Array	predicted_
	Container of the predicted output.

Array	residuals_
	Container of the residuals.

Protected Attributes inherited from STK::IRunnerSupervised< YArray_, XArray_, Weights_ >
YArray_ const *	p_y_
	A pointer on the y data set.

XArray_ const *	p_x_
	A pointer on the x data set.

Protected Attributes inherited from STK::IRunnerBase
String	msg_error_
	String with the last error message.

bool	hasRun_
	`true` if run has been used, `false` otherwise

Private Member Functions
virtual bool	regressionStep ()
	compute the regression function.

virtual bool	regressionStep (Weight const &weights)
	compute the weighted regression function.

virtual bool	predictionStep ()
	Compute the predicted outputs by the regression function.

virtual int	computeNbFreeParameter () const
	Compute the number of parameter of the regression function.

Additional Inherited Members
Protected Types inherited from STK::IRegression< Array, Array, Weight >
typedef IRunnerSupervised< Array, Array, Weight >	Base

Protected Member Functions inherited from STK::IRegression< Array, Array, Weight >
	IRegression ()
	Default constructor.

	IRegression (ArrayBase< Array > const p_y, ArrayBase< Array > const p_x)
	constructor

	IRegression (ArrayBase< Array > const &y, ArrayBase< Array > const &x)
	Constructor.

virtual bool	initializeStep ()
	perform any computation needed before the call of the regression method.

virtual bool	finalizeStep ()
	perform any computation needed after the call of the regression method.

bool	residualsStep ()
	Compute the residuals of the model.

Protected Member Functions inherited from STK::IRunnerSupervised< YArray_, XArray_, Weights_ >
	IRunnerSupervised ()
	default constructor

	IRunnerSupervised (YArray_ const const &p_y, XArray_ const const &p_x)
	constructor

	IRunnerSupervised (YArray_ const &y, XArray_ const &x)
	default constructor

	IRunnerSupervised (IRunnerSupervised const &runner)
	copy constructor

	~IRunnerSupervised ()
	destructor

virtual void	updateY ()
	update the runner when y data set is set.

virtual void	updateX ()
	update the runner when x data set is set.

virtual void	update ()
	update the runner.

Protected Member Functions inherited from STK::IRunnerBase
	IRunnerBase ()
	default constructor

	IRunnerBase (IRunnerBase const &runner)
	copy constructor

virtual	~IRunnerBase ()
	destructor

Detailed Description

template<class Array, class Weight>
class STK::MultidimRegression< Array, Weight >

The MultidimRegression class allows to regress a multidimensional output variable among a multivariate explanation variable.

Definition at line 51 of file STK_MultidimRegression.h.

Member Typedef Documentation

◆ Base

template<class Array , class Weight >

typedef IRegression<Array, Array, Weight> STK::MultidimRegression< Array, Weight >::Base

Definition at line 54 of file STK_MultidimRegression.h.

Constructor & Destructor Documentation

◆ MultidimRegression()

template<class Array , class Weight >

STK::MultidimRegression< Array, Weight >::MultidimRegression	(	Array const *	y = `0`,
		Array const *	x = `0`
	)

Constructor.

Parameters

y,x	Variates to predict and co-variates

Definition at line 92 of file STK_MultidimRegression.h.

                                                   : Base(y, x)
                                                    , coefs_()
{}

◆ ~MultidimRegression()

template<class Array , class Weight >

virtual STK::MultidimRegression< Array, Weight >::~MultidimRegression ( )

inlinevirtual

Destructor.

Definition at line 62 of file STK_MultidimRegression.h.

62{}

Member Function Documentation

◆ coefs()

template<class Array , class Weight >

Array const & STK::MultidimRegression< Array, Weight >::coefs ( ) const

inline

Returns: the coefficients

Definition at line 64 of file STK_MultidimRegression.h.

64{ return coefs_;}

References STK::MultidimRegression< Array, Weight >::coefs_.

Referenced by STK::BSplineCoefficients< Data >::extrapolate().

◆ computeNbFreeParameter()

template<class Array , class Weight >

virtual int STK::MultidimRegression< Array, Weight >::computeNbFreeParameter ( ) const

inlineprivatevirtual

Compute the number of parameter of the regression function.

Returns: the number of parameter of the regression function

Implements STK::IRegression< Array, Array, Weight >.

Definition at line 87 of file STK_MultidimRegression.h.

88 { return coefs_.sizeCols() * coefs_.sizeRows(); }

STK::IArray2D::sizeRows

int sizeRows() const

Definition STK_IArray2D.h:184

STK::IArray2D::sizeCols

int sizeCols() const

Definition STK_IArray2D.h:193

References STK::MultidimRegression< Array, Weight >::coefs_, STK::IArray2D< Derived >::sizeCols(), and STK::IArray2D< Derived >::sizeRows().

Referenced by STK::IRegression< YArray, XArray, Weights >::run(), and STK::IRegression< YArray, XArray, Weights >::run().

◆ extrapolate()

template<class Array , class Weight >

Array STK::MultidimRegression< Array, Weight >::extrapolate ( Array const & x ) const

virtual

Returns: the extrapolated values y from the value x. Given the data set x will compute the values $y = x.\hat{\beta}$ . The coefficients coefs_ have to be estimated previously.

Parameters

x	the input data set

Implements STK::IRegression< Array, Array, Weight >.

Definition at line 137 of file STK_MultidimRegression.h.

138{ return(mult(x, coefs_));}

STK::mult

Arrays::MultOp< Lhs, Rhs >::result_type mult(Lhs const &lhs, Rhs const &rhs)

convenience function for the multiplication of two matrices

Definition STK_Array2D_Functors.h:582

References STK::MultidimRegression< Array, Weight >::coefs_, and STK::mult().

Referenced by STK::AdditiveBSplineCoefficients< Array >::extrapolate().

◆ predictionStep()

template<class Array , class Weight >

bool STK::MultidimRegression< Array, Weight >::predictionStep ( )

privatevirtual

Compute the predicted outputs by the regression function.

Implements STK::IRegression< Array, Array, Weight >.

Definition at line 129 of file STK_MultidimRegression.h.

130{ this->predicted_.move(mult(*p_x_, coefs_)); return true;}

STK::IRegression< Array, Array, Weight >::predicted_

Array predicted_

Container of the predicted output.

Definition STK_IRegression.h:184

STK::IRunnerSupervised::p_x_

XArray_ const * p_x_

A pointer on the x data set.

Definition STK_IRunner.h:347

References STK::MultidimRegression< Array, Weight >::coefs_, STK::mult(), STK::IRunnerSupervised< YArray_, XArray_, Weights_ >::p_x_, and STK::IRegression< Array, Array, Weight >::predicted_.

Referenced by STK::IRegression< YArray, XArray, Weights >::run(), and STK::IRegression< YArray, XArray, Weights >::run().

◆ regressionStep() [1/2]

template<class Array , class Weight >

bool STK::MultidimRegression< Array, Weight >::regressionStep ( )

privatevirtual

compute the regression function.

Implements STK::IRegression< Array, Array, Weight >.

Definition at line 99 of file STK_MultidimRegression.h.

{
  // compute X'X
  ArraySquareX prod;
  prod.move(multLeftTranspose(p_x_->asDerived()));
  // compute (X'X)^{-1}
//  GInvertSymMatrix<ArraySquareX>()(prod);
  // compute (X'X)^{-1}X'Y
  coefs_.move(mult(invert(prod.symmetrize()), multLeftTranspose(p_x_->asDerived(), p_y_->asDerived())));
  //coefs_.move(mult(prod, multLeftTranspose(p_x_->asDerived(), p_y_->asDerived())));
  return true;
}

References STK::MultidimRegression< Array, Weight >::coefs_, STK::invert(), STK::IArray2D< Derived >::move(), STK::mult(), STK::multLeftTranspose(), STK::IRunnerSupervised< YArray_, XArray_, Weights_ >::p_x_, STK::IRunnerSupervised< YArray_, XArray_, Weights_ >::p_y_, and STK::ArrayBase< Derived >::symmetrize().

Referenced by STK::LinearAAModel< Array >::run(), STK::IRegression< YArray, XArray, Weights >::run(), STK::LinearAAModel< Array >::run(), and STK::IRegression< YArray, XArray, Weights >::run().

◆ regressionStep() [2/2]

template<class Array , class Weight >

bool STK::MultidimRegression< Array, Weight >::regressionStep ( Weight const & weights )

privatevirtual

compute the weighted regression function.

Parameters

weights the weights of the samples

Implements STK::IRegression< Array, Array, Weight >.

Definition at line 114 of file STK_MultidimRegression.h.

{
  // compute X'WX
  ArraySquareX prod;
  prod.move(weightedMultLeftTranspose(p_x_->asDerived(), weights));
  // compute (X'WX)^{-1}
  //GInvertSymMatrix<ArraySquareX>()(prod);
  // compute (X'WX)^{-1}X'WY
  coefs_.move(mult(invert(prod.symmetrize()), wmultLeftTranspose(p_x_->asDerived(), p_y_->asDerived(), weights)));
//  coefs_.move(mult(prod, wmultLeftTranspose(p_x_->asDerived(), p_y_->asDerived(), weights)));
  return true;
}

References STK::MultidimRegression< Array, Weight >::coefs_, STK::invert(), STK::IArray2D< Derived >::move(), STK::mult(), STK::IRunnerSupervised< YArray_, XArray_, Weights_ >::p_x_, STK::IRunnerSupervised< YArray_, XArray_, Weights_ >::p_y_, STK::ArrayBase< Derived >::symmetrize(), STK::weightedMultLeftTranspose(), and STK::wmultLeftTranspose().

Member Data Documentation

◆ coefs_

template<class Array , class Weight >

ArrayXX STK::MultidimRegression< Array, Weight >::coefs_

protected

Definition at line 73 of file STK_MultidimRegression.h.

Referenced by STK::MultidimRegression< Array, Weight >::coefs(), STK::MultidimRegression< Array, Weight >::computeNbFreeParameter(), STK::AdditiveBSplineCoefficients< Array >::extrapolate(), STK::MultidimRegression< Array, Weight >::extrapolate(), STK::AdditiveBSplineRegression< YArray, XArray, Weights >::extrapolate(), STK::BSplineRegression< YArray, XVector, Weights >::extrapolate(), STK::AdditiveBSplineRegression< YArray, XArray, Weights >::initializeStep(), STK::BSplineRegression< YArray, XVector, Weights >::initializeStep(), STK::AdditiveBSplineRegression< YArray, XArray, Weights >::predictionStep(), STK::BSplineRegression< YArray, XVector, Weights >::predictionStep(), STK::MultidimRegression< Array, Weight >::predictionStep(), STK::AdditiveBSplineRegression< YArray, XArray, Weights >::regressionStep(), STK::BSplineRegression< YArray, XVector, Weights >::regressionStep(), STK::MultidimRegression< Array, Weight >::regressionStep(), STK::MultidimRegression< Array, Weight >::regressionStep(), STK::AdditiveBSplineRegression< YArray, XArray, Weights >::regressionStep(), STK::BSplineRegression< YArray, XVector, Weights >::regressionStep(), and STK::AdditiveBSplineCoefficients< Array >::run().

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

STK_MultidimRegression.h

Public Types

Public Member Functions

Protected Attributes

Private Member Functions

Additional Inherited Members

Detailed Description

Member Typedef Documentation

◆ Base

Constructor & Destructor Documentation

◆ MultidimRegression()

◆ ~MultidimRegression()

Member Function Documentation

◆ coefs()

◆ computeNbFreeParameter()

◆ extrapolate()

◆ predictionStep()

◆ regressionStep() [1/2]

◆ regressionStep() [2/2]

Member Data Documentation

◆ coefs_