jhplot.bsom

Class Matrix

java.lang.Object

jhplot.bsom.Matrix

Direct Known Subclasses:

UserData

public class Matrixextends Object

Matrix is a class for matrix objects with basic linear calculation methods.

Constructor Summary

Constructors

Constructor and Description
Matrix(int n) Construct a square matrix object.
Matrix(int row, int col) Construct a rectangle matrix object.

Method Summary

Modifier and Type	Method and Description
Matrix	copy() Copy the matrix.
Matrix	crossSqDistance(Matrix Y) Cross squared Euclidean distance Matrix: Z_{ij} = ||X_i - Y_j||^2
Matrix	diagonal() If the matrix is a row or column vector, make a diagonal matrix from it.
Matrix	dividedBy(Matrix Y) Matrix division: X\Y.
Matrix	dividedBy(Matrix Y, int bw) Matrix division with bandwidth: X\Y.
Matrix	eigenvalues(double eps, int lmax) Make a row vector including the eigenvalues of the matrix.
Matrix	horizontalSum() Make a horizontal-summation vector.
static Matrix	identity(int n) Make an identity matrix.
Matrix	linearConv(double a, double b, Matrix Y) Linear conversion: a*X + b*Y.
Matrix	mulipliedEntriesWith(Matrix Y) Multiply entries.
Matrix	multipliedBy(double a) Scalar product: aX.
Matrix	multipliedBy(Matrix Y) Matrix multiplication: X*Y.
static Matrix	random(int row, int col) Same as random(row, col, 1).
static Matrix	random(int row, int col, int seed) Make a standard Gaussian random Matrix.
double	sumEntries() Summation of all entries.
double	sumSqrEntries() Summation of squared entries.
Matrix	transpose() Transpose: X'.
void	updateDivideByRowVector(Matrix vec) X_{ij} = X_{ij}/v_{j}.
void	updateExp() Update by exponential : exp(X_{ij})
void	updateLinearConv(double a, double b, Matrix Y) Update by Linear conversion: X = a*X + b*Y.
Matrix	verticalSum() Make a vertical-summation vector.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Matrix

public Matrix(int n)

Construct a square matrix object.

Parameters:

n - size

Matrix

public Matrix(int row,              int col)

Construct a rectangle matrix object.

Parameters:

row - number of row

col - number of column

Method Detail

identity

public static Matrix identity(int n)

Make an identity matrix.

Parameters:

n - size

random

public static Matrix random(int row,                            int col,                            int seed)

Make a standard Gaussian random Matrix.

Parameters:

row - number of row

col - number of column

seed - random seed

random

public static Matrix random(int row,                            int col)

Same as random(row, col, 1).

copy

public Matrix copy()

Copy the matrix.

transpose

public Matrix transpose()

Transpose: X'.

linearConv

public Matrix linearConv(double a,                         double b,                         Matrix Y)

Linear conversion: a*X + b*Y.

updateLinearConv

public void updateLinearConv(double a,                             double b,                             Matrix Y)

Update by Linear conversion: X = a*X + b*Y.

multipliedBy

public Matrix multipliedBy(double a)

Scalar product: aX.

multipliedBy

public Matrix multipliedBy(Matrix Y)

Matrix multiplication: X*Y.

dividedBy

public Matrix dividedBy(Matrix Y)

Matrix division: X\Y.

dividedBy

public Matrix dividedBy(Matrix Y,                        int bw)

Matrix division with bandwidth: X\Y.

crossSqDistance

public Matrix crossSqDistance(Matrix Y)

Cross squared Euclidean distance Matrix: Z_{ij} = ||X_i - Y_j||^2

updateExp

public void updateExp()

Update by exponential : exp(X_{ij})

horizontalSum

public Matrix horizontalSum()

Make a horizontal-summation vector.

verticalSum

public Matrix verticalSum()

Make a vertical-summation vector.

sumEntries

public double sumEntries()

Summation of all entries.

sumSqrEntries

public double sumSqrEntries()

Summation of squared entries.

mulipliedEntriesWith

public Matrix mulipliedEntriesWith(Matrix Y)

Multiply entries.

updateDivideByRowVector

public void updateDivideByRowVector(Matrix vec)

X_{ij} = X_{ij}/v_{j}.

diagonal

public Matrix diagonal()

If the matrix is a row or column vector, make a diagonal matrix from it. Otherwise extract its diagonal as a row vector.

eigenvalues

public Matrix eigenvalues(double eps,                          int lmax)

Make a row vector including the eigenvalues of the matrix.

Parameters:

eps - allowable absolute value of off-diagonal entries