LTSA Class Reference

#include <ltsa.h>

Inheritance diagram for LTSA:

Collaboration diagram for LTSA:

Public Member Functions
void	setSpecificParameters (int k=12)
	Set specific parameters. More...
virtual void	reduceDimensionality ()
	Reduce dimensionality. More...
Public Member Functions inherited from DimRedAlgorithm
	DimRedAlgorithm ()
	Empty constructor. More...
void	setInputData (Matrix2D< double > &X)
	Set input data. More...
void	setOutputDimensionality (size_t outputDim)
	Set output dimensionality. More...
const Matrix2D< double > &	getReducedData ()
	Get reduced data. More...

Public Attributes
int	k
Public Attributes inherited from DimRedAlgorithm
Matrix2D< double > *	X
	Pointer to input data. More...
size_t	outputDim
	Output dim. More...
Matrix2D< double >	Y
	Output data. More...
DimRedDistance2	distance
	Distance function. More...
FileName	fnMapping
	Save mapping. More...

Protected Member Functions
void	computeAlignmentMatrix (Matrix2D< double > &B)
	Common part. More...

Detailed Description

Class for making a LTSA dimensionality reduction

Definition at line 36 of file ltsa.h.

Member Function Documentation

computeAlignmentMatrix()

void LTSA::computeAlignmentMatrix ( Matrix2D< double > &  B )

protected

Common part.

Definition at line 56 of file ltsa.cpp.

{
    subtractColumnMeans(*X);

    size_t n = MAT_YSIZE(*X);
    Matrix2D<int> ni;
    Matrix2D<double> D;
    kNearestNeighbours(*X, k, ni, D);
    Matrix2D<double> Xi(MAT_XSIZE(ni), MAT_XSIZE(*X)), W, Vi, Vi2, Si, Gi;

    B.initIdentity(n);
    Matrix1D<int> weightVector;
    for (size_t iLoop = 0; iLoop < n; ++iLoop)
    {
        extractNearestNeighbours(*X, ni, iLoop, Xi);
        subtractColumnMeans(Xi);

        matrixOperation_AAt(Xi, W); // W=X*X^t
        schur(W, Vi, Si);           // W=Vi*Si*Vi^t

        computeWeightsVector(Si, weightVector);

        Vi2.resizeNoCopy(MAT_YSIZE(Vi), outputDim + 1);
        Vi2.setConstantCol(0, 1/sqrt(k)); //Vi2(0,:)=1/sqrt(k)
        getLessWeightNColumns(Vi, weightVector, Vi2);

        matrixOperation_AAt(Vi2, Gi); // Gi=Vi2*Vi2^t
        matrixOperation_IminusA(Gi);  // Gi=I-Gi

        // Compute partial B with correlation matrix Gi
        FOR_ALL_ELEMENTS_IN_MATRIX2D(Gi)
            MAT_ELEM(B, MAT_ELEM(ni,iLoop,i), MAT_ELEM(ni,iLoop,j)) += MAT_ELEM(Gi, i, j);
        MAT_ELEM(B, iLoop, iLoop)-=1;
    }
}

reduceDimensionality()

void LTSA::reduceDimensionality ( )

virtual

Reduce dimensionality.

Implements DimRedAlgorithm.

Reimplemented in LLTSA.

Definition at line 92 of file ltsa.cpp.

{
    Matrix2D<double> B;
    computeAlignmentMatrix(B);

    Matrix1D<double> DEigs;
    eigsBetween(B, 1, outputDim, DEigs, Y);
}

setSpecificParameters()

void LTSA::setSpecificParameters

(

int

k = 12

)

Set specific parameters.

Definition at line 26 of file ltsa.cpp.

{
    this->k = k;
}

Member Data Documentation

k

int LTSA::k

Definition at line 39 of file ltsa.h.

---

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

• xmipp/libraries/dimred/ltsa.h
• xmipp/libraries/dimred/ltsa.cpp