com.tagtraum.jipes.math

Interface Matrix

All Known Subinterfaces:

MutableMatrix

All Known Implementing Classes:

AbstractMatrix, FullMatrix, MutableAbstractMatrix, SparseColumnMatrix, SparseMatrix, SparseRowMatrix, SymmetricBandMatrix, SymmetricMatrix

public interface Matrix

Matrix. All mathematical operations are lazy—the resulting matrices are only views still referencing their operands. I.e., If the operands are mutable (MutableMatrix), they results change dynamically.

Author:

Hendrik Schreiber

Method Summary

Modifier and Type	Method and Description
Matrix	add(Matrix m) Creates a view that is equal to the result of an addition.
float[]	columnSum() Compute the sum for each column.
Matrix	enlarge(Matrix m) Creates a view that is as large as the larger of the two involved matrices.
float	get(int row, int column) Gets the value from row and column.
float[]	getColumn(int column) Gets the value from column.
int	getNumberOfColumns() Number of columns.
int	getNumberOfRows() Number of rows
float[]	getRow(int row) Gets the values from row.
Matrix	hadamardMultiply(Matrix m) Creates a view that is equal to the result of a Hadamard multiplication.
boolean	isZeroPadded() Indicates whether this matrix is zero padded.
Matrix	multiply(float f) Creates a view that is equal to the result of a scalar multiplication.
Matrix	multiply(Matrix m) Creates a view that is equal to the result of a matrix multiplication.
float[]	rowSum() Compute the sum for each row.
Matrix	subtract(Matrix m) Creates a view that is equal to the result of an subtraction.
float	sum() Compute the sum of all elements.
Matrix	translate(int rows, int columns) Creates a view that is equal to the result of a matrix translation.
Matrix	transpose() Creates a view that is equal to the result of a matrix transposition.

Method Detail

get

float get(int row,
          int column)

Gets the value from row and column. If the matrix supports zero-padding, this method may return 0f for undefined rows or columns.

Parameters:

row - row

column - column

Returns:

value

Throws:

IndexOutOfBoundsException - if row, column or both are out of bounds and the matrix does not support zero padding

sum

float sum()

Compute the sum of all elements.

Returns:

sum

rowSum

float[] rowSum()

Compute the sum for each row.

Returns:

sum for all rows

columnSum

float[] columnSum()

Compute the sum for each column.

Returns:

sum for all columns

getRow

float[] getRow(int row)

Gets the values from row. If the matrix supports zero-padding, this method may return 0f for undefined rows or columns.

Parameters:

row - row

Returns:

value

Throws:

IndexOutOfBoundsException - if the row is out of bounds and the matrix does not support zero padding

getColumn

float[] getColumn(int column)

Gets the value from column. If the matrix supports zero-padding, this method may return 0f for undefined columns.

Parameters:

column - column

Returns:

value

Throws:

IndexOutOfBoundsException - if the column is out of bounds and the matrix does not support zero padding

getNumberOfRows

int getNumberOfRows()

Number of rows

Returns:

rows

getNumberOfColumns

int getNumberOfColumns()

Number of columns.

Returns:

columns

isZeroPadded

boolean isZeroPadded()

Indicates whether this matrix is zero padded. I.e. access to undefined rows or columns leads to a return value 0f, instead of an exception.

Returns:

true or false

add

Matrix add(Matrix m)

Creates a view that is equal to the result of an addition.

Parameters:

m - matrix to add to this

Returns:

add view

subtract

Matrix subtract(Matrix m)

Creates a view that is equal to the result of an subtraction.

Parameters:

m - matrix to subtract from this

Returns:

subtract view

multiply

Matrix multiply(float f)

Creates a view that is equal to the result of a scalar multiplication.

Parameters:

f - scalar

Returns:

multiplication view

multiply

Matrix multiply(Matrix m)

Creates a view that is equal to the result of a matrix multiplication.

Parameters:

m - matrix to multiply this with

Returns:

multiplication view

hadamardMultiply

Matrix hadamardMultiply(Matrix m)

Creates a view that is equal to the result of a Hadamard multiplication.

Parameters:

m - matrix to multiply this with

Returns:

multiplication view

transpose

Matrix transpose()

Creates a view that is equal to the result of a matrix transposition.

Returns:

transposition view

enlarge

Matrix enlarge(Matrix m)

Creates a view that is as large as the larger of the two involved matrices. Values are equal to the values of this matrix, as long as its maximal indices aren't exceeded. Once they are exceeded, the given matrix is used as "backup". If its indices are also exceeded its padding (if available) is returned upon read operations.
In essence this operation adds the given matrix to this matrix for those rows and columns where this matrix isn't defined. The existing values of the given matrix are invisible for those indices where this matrix is defined.

Parameters:

m - matrix

Returns:

enlarged matrix combining this and the given matrix

translate

Matrix translate(int rows,
                 int columns)

Creates a view that is equal to the result of a matrix translation. I.e. the content of the matrix is moved to a different position in a larger matrix.

Parameters:

rows - number of rows to move the matrix

columns - number of columns to move the matrix

Returns:

transposition view