Class holding a Matrix that can be used to decorrelate input data to Machine Learning classifiers. More...

#include <DecorrelationMatrix.h>

Public Types
using	MatrixT = Eigen::Matrix< double, Ndims, Ndims, Eigen::RowMajor >
	typedef for consistent type across class

Public Member Functions
	DecorrelationMatrix (const MatrixT &matrix=MatrixT::Identity())
	default constructor: initializes to identity matrix or to passed matrix

	DecorrelationMatrix (const DecorrelationMatrix &matrix)=default
	copy constructor

DecorrelationMatrix &	operator= (const DecorrelationMatrix &rhs)=default
	assignment operator

const MatrixT &	getMatrix () const
	get the currently stored matrix

void	calculateDecorrMatrix (std::array< std::vector< double >, Ndims > inputData, bool normalise=true)
	calculate the transformation matrix that when applied to the input data yields linearly uncorrelated data.

void	calculateDecorrMatrix (const MatrixT &covMatrix, bool normalise=true)
	calculate the matrix that can be used to decorrelate the data that yield the passed covariance matrix.

std::vector< double >	decorrelate (const std::vector< double > &inputVec) const
	"decorrelate" one measurement (i.e.

std::vector< double >	decorrelate (const std::array< double, Ndims > &inputs) const
	"decorrelate" one measurement (i.e.

std::array< std::vector< double >, Ndims >	decorrelate (const std::array< std::vector< double >, Ndims > &inputMat) const
	decorrelate M measurements (i.e.

std::string	print () const
	print the matrix to a string.

bool	readFromStream (std::istream &is)
	read from stream.

Private Attributes
MatrixT	m_matrix
	internal matrix holding the data

Detailed Description

template<size_t Ndims>
class Belle2::DecorrelationMatrix< Ndims >

Class holding a Matrix that can be used to decorrelate input data to Machine Learning classifiers.

At the moment only the dimension is templated and defaults to Ndims = 9, as it will be mainly used for ThreeHitFilters which have at the moment only 9 inputs.

Definition at line 36 of file DecorrelationMatrix.h.

Member Typedef Documentation

◆ MatrixT

using MatrixT = Eigen::Matrix<double, Ndims, Ndims, Eigen::RowMajor>

typedef for consistent type across class

Definition at line 39 of file DecorrelationMatrix.h.

Constructor & Destructor Documentation

◆ DecorrelationMatrix()

DecorrelationMatrix ( const MatrixT & matrix = MatrixT::Identity() )

inlineexplicit

default constructor: initializes to identity matrix or to passed matrix

Definition at line 42 of file DecorrelationMatrix.h.

42: m_matrix(matrix) { }

Belle2::DecorrelationMatrix::m_matrix

MatrixT m_matrix

internal matrix holding the data

Definition: DecorrelationMatrix.h:89

Member Function Documentation

◆ getMatrix()

const MatrixT & getMatrix ( ) const

inline

get the currently stored matrix

Definition at line 50 of file DecorrelationMatrix.h.

Member Data Documentation

◆ m_matrix

MatrixT m_matrix

private

internal matrix holding the data

Definition at line 89 of file DecorrelationMatrix.h.

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

tracking/trackFindingVXD/filterTools/include/DecorrelationMatrix.h

Public Types

Public Member Functions

Private Attributes