lti::linearRegression< T > Class Template Reference
[Linear Algebra]

The linear regression algorithm estimates a linear relation between two vector spaces. More...

#include <ltiLinearRegression.h>

Inheritance diagram for lti::linearRegression< T >:

[legend]

Collaboration diagram for lti::linearRegression< T >:

[legend]

List of all members.

Classes
class	parameters
	the parameters for the class linearRegression More...
Public Member Functions
	linearRegression ()
	linearRegression (const parameters &par)
	linearRegression (const linearRegression &other)
virtual	~linearRegression ()
virtual const char *	getTypeName () const
matrix< T >	apply (matrix< T > &srcMatrix, matrix< T > &destMatrix)
void	transform (vector< T > &srcVector, vector< T > &destVector) const
vector< T >	transform (vector< T > &srcVector) const
matrix< T >	getLinRegMatrix () const
void	setLinRegMatrix (matrix< T > &linRegMatrixSet)
linearRegression &	copy (const linearRegression &other)
linearRegression &	operator= (const linearRegression &other)
virtual functor *	clone () const
const parameters &	getParameters () const
virtual bool	read (ioHandler &handler, const bool complete=true)
virtual bool	write (ioHandler &handler, const bool complete=true) const

---

Detailed Description

template<class T>
class lti::linearRegression< T >

The linear regression algorithm estimates a linear relation between two vector spaces.

To each vector srcVect of the source vector space with dimension srcDim, the corresponding vector destDim of the destination vector space with dimension destDim can be calculated using the resulting regression matrix R:

destVect = R * srcVect.

R is calculated from a training set of corresponding vectors in both spaces. Note that srcVect = 0 can only result in destVect = 0, so it is important to eliminate any offsets from both vector spaces before using this algorithm:

destVect = destVectMean + R * (srcVect - srcVectMean)

For example application see ltiLinearRegressionTracking.

---

Constructor & Destructor Documentation

template<class T>

lti::linearRegression< T >::linearRegression

(

)

default constructor

template<class T>

lti::linearRegression< T >::linearRegression

(

const parameters &

par

)

default constructor with parameters

template<class T>

lti::linearRegression< T >::linearRegression

(

const linearRegression< T > &

other

)

copy constructor

Parameters:

other

the object to be copied

template<class T>

virtual lti::linearRegression< T >::~linearRegression

(

)

[virtual]

destructor

---

Member Function Documentation

template<class T>

matrix<T> lti::linearRegression< T >::apply	(	matrix< T > &	srcMatrix,
		matrix< T > &	destMatrix
	)

The linear regression algorithm.

It calculates a linear transformation matrix linRegMatrix, that approximately fulfills destMatrix = linRegMatrix * srcMatrix.

Parameters:

	srcMatrix	the columns of this matrix are the source training vectors of dimension srcDim
	destMatrix	the columns of this matrix are the destination training vectors of dimension destDim

Returns:

resulting (destDim x srcDim) transformation matrix. This is also kept in the parameter linRegMatrix and can be used directly with "transform"

template<class T>

virtual functor* lti::linearRegression< T >::clone

(

)

const [virtual]

returns a pointer to a clone of this functor.

Implements lti::functor.

Reimplemented in lti::linearRegressionTracking.

template<class T>

linearRegression& lti::linearRegression< T >::copy

(

const linearRegression< T > &

other

)

copy data of "other" functor.

Parameters:

other

the functor to be copied

Returns:

a reference to this functor object

Reimplemented from lti::functor.

template<class T>

matrix<T> lti::linearRegression< T >::getLinRegMatrix

(

)

const

gets the linear regression matrix

template<class T>

const parameters& lti::linearRegression< T >::getParameters

(

)

const

returns used parameters

Reimplemented from lti::functor.

Reimplemented in lti::linearRegressionTracking.

template<class T>

virtual const char* lti::linearRegression< T >::getTypeName

(

)

const [virtual]

returns the name of this type ("linearRegression")

Reimplemented from lti::linearAlgebraFunctor.

Reimplemented in lti::linearRegressionTracking.

template<class T>

linearRegression& lti::linearRegression< T >::operator=

(

const linearRegression< T > &

other

)

Alias for copy.

Reimplemented from lti::functor.

template<class T>

virtual bool lti::linearRegression< T >::read	(	ioHandler &	handler,
		const bool	complete = true
	)			[virtual]

Reads this functor from the given handler.

Reimplemented from lti::functor.

Reimplemented in lti::linearRegressionTracking.

template<class T>

void lti::linearRegression< T >::setLinRegMatrix

(

matrix< T > &

linRegMatrixSet

)

sets the linear regression matrix

template<class T>

vector<T> lti::linearRegression< T >::transform

(

vector< T > &

srcVector

)

const

multiplication of srcVector with the linear regression matrix of dimension (destDim x srcDim)

Parameters:

srcVector

source vector of dimension srcDim

Returns:

destination vector of dimension destDim

template<class T>

void lti::linearRegression< T >::transform	(	vector< T > &	srcVector,
		vector< T > &	destVector
	)			const

multiplication of srcVector with the linear regression matrix of dimension (destDim x srcDim)

Parameters:

	srcVector	source vector of dimension srcDim
	destVector	destination vector of dimension destDim

template<class T>

virtual bool lti::linearRegression< T >::write	(	ioHandler &	handler,
		const bool	complete = true
	)			const [virtual]

Writes this functor to the given handler.

Reimplemented from lti::functor.

Reimplemented in lti::linearRegressionTracking.

---

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

• ltiLinearRegression.h