mikera.matrixx.impl

Class SparseRowMatrix

java.lang.Object

mikera.arrayz.impl.AbstractArray<AVector>

mikera.matrixx.AMatrix

mikera.matrixx.impl.ARectangularMatrix

mikera.matrixx.impl.ASparseRCMatrix

mikera.matrixx.impl.SparseRowMatrix

All Implemented Interfaces:

java.io.Serializable, java.lang.Cloneable, java.lang.Comparable<INDArray>, java.lang.Iterable<AVector>, INDArray, ISparse, IMatrix, IFastRows

public class SparseRowMatrix
extends ASparseRCMatrix
implements ISparse, IFastRows

Matrix stored as a collection of normally sparse row vectors. This format is especially efficient for: - innerProduct() with another matrix, especially one with efficient column access like SparseColumnMatrix - access via getRow() operation - transpose into SparseColumnMatrix

See Also:

Serialized Form

Method Summary

Modifier and Type	Method and Description
void	add(AMatrix a) Adds another matrix to this matrix.
void	addAt(int i, int j, double d) Adds a value at a specific position in the matrix, mutating the matrix Does not perform bounds checking - this in an unsafe operation
void	addToArray(double[] targetData, int offset) Adds all the elements of this array to a double array at the specified offset, in row-major order
AMatrix	clone() Returns a clone of the array data, as a new array which will be fully mutable and may be of a different class to the original.
int	componentCount() Returns the number of components of this array.
void	copyColumnTo(int col, double[] targetData, int offset) Copies the elements in a selected row of this matrix to a double array
void	copyRowTo(int i, double[] data, int offset) Copies the elements in a selected row of this matrix to a double array
static SparseRowMatrix	create(AMatrix source)
static INDArray	create(java.util.ArrayList<INDArray> slices, int rows, int cols)
static SparseRowMatrix	create(AVector... vecs)
static SparseRowMatrix	create(AVector[] data, int rows, int cols)
static SparseRowMatrix	create(int rows, int cols)
static SparseRowMatrix	create(java.util.List<AVector> rows)
boolean	equals(AMatrix m) Returns true if this matrix is exactly equal to another matrix
boolean	equals(IFastRows m) Compares this matrix to another matrix implementing IFastColumns This can be much faster than the default equals
SparseRowMatrix	exactClone() Creates a deep clone of an array, using the same class implementation as the original array
double	get(int i, int j) Returns the double value at the specified position in a 2D matrix
java.util.List<AVector>	getColumns() Gets a List of columns of a matrix.
AVector	getComponent(int k) Gets a component from the array at the specified index Will throw an error if components are not supported, or if the component is out of bounds as defined by 0 <= k
AVector	getRow(int i) Returns a row of the matrix.
java.util.List<AVector>	getRows() Gets a List of rows of a matrix.
AVector	getRowView(int i) Returns a row of the matrix as a vector view.
SparseColumnMatrix	getTransposeView() Returns a transposed view of the array.
AMatrix	innerProduct(AMatrix a) Computes the inner product of this matrix with another.
AVector	innerProduct(AVector a) Computes the inner product of this array with a vector
SparseRowMatrix	innerProduct(SparseColumnMatrix a) Specialised inner product for sparse row matrix multiplied by sparse column matrix.
SparseRowMatrix	innerProduct(SparseRowMatrix a)
boolean	isUpperTriangular() Returns true if a matrix is upper triangular An upper triangular matrix is defined as having all elements equal to 0.0 where i > j
AMatrix	multiplyCopy(double a) Returns a copy of the array with all elements multiplied by a single constant value
void	multiplyRow(int i, double value) Multiplies a row by a constant factor This is an elementary row operation
void	replaceRow(int i, AVector vec) Replaces a row in a matrix, adding the row to the internal structure of the matrix.
double	rowDotProduct(int i, AVector a) Returns the dot product of a specific row with a vector.
void	set(AMatrix a) Sets this matrix with the element values from another matrix.
void	set(int i, int j, double value) Sets a value at a given position in a mutable 2D array
void	setRow(int i, AVector v) Sets a row in a matrix to the value specified by the given vector
AMatrix	sparse() Coerces this INDArray to a sparse format, without changing its element values.
void	swapRows(int i, int j) Swaps two rows of the matrix in place This is an elementary row operation
double[]	toDoubleArray() Copies the elements of this array to a new double[] array.
SparseColumnMatrix	toSparseColumnMatrix()
double	unsafeGet(int row, int column) Gets an element in the matrix in an unsafe fashion, without performing bound checks The result is undefined if the row and column are out of bounds.
void	unsafeSet(int i, int j, double value) Sets an element value in the matrix in an unsafe fashion, without performing bound checks The result is undefined if the row and column are out of bounds.
void	validate() Validates the internal data structure of the INDArray.
static SparseRowMatrix	wrap(AVector... vecs)
static SparseRowMatrix	wrap(AVector[] vecs, int rows, int cols)
static SparseRowMatrix	wrap(java.util.List<AVector> vecs)

Methods inherited from class mikera.matrixx.impl.ASparseRCMatrix

abs, applyOp, dense, elementMax, elementMin, elementSquaredSum, elementSum, exp, fill, getColumnClone, getRowClone, isFullyMutable, isMutable, isSparse, isZero, log, nonZeroCount, pow, reciprocal, setSparse, sqrt, square, unsafeGetVector

Methods inherited from class mikera.matrixx.impl.ARectangularMatrix

bandLength, checkColumn, checkRow, checkSquare, columnCount, elementCount, getElement, getShape, getShape, getShapeClone, isSameShape, isSameShape, isSameShape, isSquare, rowCount

Methods inherited from class mikera.matrixx.AMatrix

add, add, add, add2, addCopy, addCopy, addCopy, addCopy, addCopy, addInnerProduct, addInnerProduct, addInnerProduct, addMultiple, addMultiple, addMultiple, addOuterProduct, addOuterProduct, addOuterProductSparse, addOuterProductSparse, addRowMultiple, addSparse, addSparse, applyOp, applyOp, applyOp, applyOp, applyOp, applyOpCopy, asDoubleArray, asElementList, asVector, bandIndex, bandPosition, bandStartColumn, bandStartRow, broadcast, broadcastCloneLike, broadcastCopyLike, broadcastLike, broadcastLike, clamp, composeWith, copy, denseClone, determinant, diagonalProduct, dimensionality, divide, divide, divide, divide, divideCopy, elementIterator, elementsEqual, ensureMutable, epsilonEquals, epsilonEquals, epsilonEquals, equals, equals, equalsArray, equalsTranspose, get, get, get, get, getBand, getBandWrapped, getColumn, getColumnView, getElements, getLeadingDiagonal, getLongShape, getSlices, getSlices, getSliceViews, getTranspose, getTransposeCopy, hashCode, hasOrthonormalColumns, hasOrthonormalRows, hasUncountable, immutable, innerProduct, innerProduct, innerProduct, innerProduct, inputDimensions, inverse, isBoolean, isDiagonal, isElementConstrained, isHermitian, isIdentity, isInvertible, isLowerTriangular, isOrthogonal, isOrthogonal, isPositiveDefinite, isRectangularDiagonal, isSymmetric, isView, iterator, join, lowerBandwidth, lowerBandwidthLimit, mul, multiply, multiply, multiply, multiply, mutable, negate, outerProduct, outputDimensions, rank, reduce, reduce, reduceSlices, reduceSlices, reorder, reorder, replaceColumn, reshape, reshape, rotateView, scaleAdd, scaleAdd, set, set, set, set, set, set, set, setApplyOp, setApplyOp, setColumn, setElements, setElements, setMultiple, setMultiple, setMultiple, setMultiple, setMultiple, setSparse, setSparse, signum, slice, slice, sliceCount, sparseClone, sub, sub, sub, subArray, subCopy, subMatrix, swapColumns, toAffineTransform, toDoubleBuffer, toMatrix, toMatrixTranspose, toMutableMatrix, toNestedDoubleArrays, toSliceArray, toString, toStringFull, toVector, trace, transform, transform, transformInPlace, transformInPlace, transposeInnerProduct, transposeInnerProduct, transposeInPlace, upperBandwidth, upperBandwidthLimit

Methods inherited from class mikera.arrayz.impl.AbstractArray

absCopy, absDiff, absDiffCopy, add, addApplyOp, addAt, addCopy, addInnerProduct, addInnerProduct, addMultipleSparse, addPower, addPower, addSparse, broadcastLike, compareTo, density, divideCopy, elementAbsPowSum, elementMaxAbs, elementPowSum, elementProduct, epsilonEquals, equals, equalsArray, get, get, get, get, getComponents, getElements, getElements, getElements, isDense, join, multiplyCopy, negateCopy, reciprocalCopy, scale, scaleCopy, setElements, setInnerProduct, signumCopy, sliceValue, squareCopy, sub, sub, subCopy, withComponents

Methods inherited from class java.lang.Object

getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface mikera.arrayz.ISparse

density

Methods inherited from interface mikera.matrixx.IMatrix

add2, addCopy, addCopy, columnCount, diagonalProduct, getBand, getColumn, getColumnClone, getColumnView, getRowClone, getTranspose, innerProduct, inverse, isDiagonal, isIdentity, isInvertible, isRectangularDiagonal, isSquare, isSymmetric, rowCount, subMatrix, toNestedDoubleArrays, trace, transform, transformInPlace, transposeInPlace

Methods inherited from interface mikera.arrayz.INDArray

abs, absCopy, absDiff, absDiffCopy, add, add, addApplyOp, addAt, addCopy, addCopy, addInnerProduct, addInnerProduct, addMultiple, addMultipleSparse, addOuterProduct, addOuterProductSparse, addPower, addPower, addSparse, addSparse, applyOp, applyOp, applyOp, applyOp, applyOpCopy, asDoubleArray, asElementList, asVector, broadcast, broadcastCloneLike, broadcastCopyLike, broadcastLike, broadcastLike, broadcastLike, clamp, copy, dense, denseClone, dimensionality, divide, divide, divideCopy, divideCopy, elementAbsPowSum, elementCount, elementIterator, elementMax, elementMaxAbs, elementMin, elementPowSum, elementProduct, elementsEqual, elementSquaredSum, elementSum, ensureMutable, epsilonEquals, epsilonEquals, equals, equalsArray, equalsArray, exp, fill, get, get, get, get, get, get, get, get, getComponents, getElement, getElements, getElements, getElements, getElements, getLongShape, getShape, getShape, getShapeClone, getSlices, getSlices, getSliceViews, getTransposeCopy, hasUncountable, immutable, innerProduct, innerProduct, innerProduct, isBoolean, isDense, isElementConstrained, isFullyMutable, isMutable, isSameShape, isSparse, isView, isZero, join, join, log, multiply, multiply, multiplyCopy, mutable, negate, negateCopy, nonZeroCount, outerProduct, pow, reciprocal, reciprocalCopy, reduce, reduce, reduceSlices, reduceSlices, reorder, reorder, reshape, rotateView, scale, scaleAdd, scaleAdd, scaleCopy, set, set, set, set, set, set, setApplyOp, setElements, setElements, setElements, setInnerProduct, setMultiple, setMultiple, setSparse, setSparse, signum, signumCopy, slice, slice, sliceCount, sliceValue, sparseClone, sqrt, square, squareCopy, sub, sub, subArray, subCopy, toDoubleBuffer, toSliceArray, toVector, withComponents

Methods inherited from interface java.lang.Comparable

compareTo

Methods inherited from interface java.lang.Iterable

forEach, spliterator

Method Detail

create

public static SparseRowMatrix create(int rows,
                                     int cols)

create

public static SparseRowMatrix create(AVector[] data,
                                     int rows,
                                     int cols)

create

public static SparseRowMatrix create(AVector... vecs)

create

public static SparseRowMatrix create(java.util.List<AVector> rows)

create

public static INDArray create(java.util.ArrayList<INDArray> slices,
                              int rows,
                              int cols)

wrap

public static SparseRowMatrix wrap(AVector[] vecs,
                                   int rows,
                                   int cols)

wrap

public static SparseRowMatrix wrap(AVector... vecs)

create

public static SparseRowMatrix create(AMatrix source)

wrap

public static SparseRowMatrix wrap(java.util.List<AVector> vecs)

componentCount

public int componentCount()

Description copied from interface: INDArray

Returns the number of components of this array. May return 0 if components are not supported

Specified by:

componentCount in interface INDArray

Specified by:

componentCount in class ASparseRCMatrix

Returns:

getComponent

public AVector getComponent(int k)

Description copied from interface: INDArray

Gets a component from the array at the specified index Will throw an error if components are not supported, or if the component is out of bounds as defined by 0 <= k

Specified by:

getComponent in interface INDArray

Specified by:

getComponent in class ASparseRCMatrix

Returns:

get

public double get(int i,
                  int j)

Description copied from interface: INDArray

Returns the double value at the specified position in a 2D matrix

Specified by:

get in interface INDArray

Specified by:

get in interface IMatrix

Specified by:

get in class AMatrix

set

public void set(int i,
                int j,
                double value)

Description copied from interface: INDArray

Sets a value at a given position in a mutable 2D array

Specified by:

set in interface INDArray

Specified by:

set in interface IMatrix

Overrides:

set in class ARectangularMatrix

unsafeGet

public double unsafeGet(int row,
                        int column)

Description copied from class: AMatrix

Gets an element in the matrix in an unsafe fashion, without performing bound checks The result is undefined if the row and column are out of bounds.

Overrides:

unsafeGet in class AMatrix

Returns:

unsafeSet

public void unsafeSet(int i,
                      int j,
                      double value)

Description copied from class: AMatrix

Sets an element value in the matrix in an unsafe fashion, without performing bound checks The result is undefined if the row and column are out of bounds.

Overrides:

unsafeSet in class AMatrix

set

public void set(AMatrix a)

Description copied from class: AMatrix

Sets this matrix with the element values from another matrix. Source matrix must have the same shape.

Overrides:

set in class AMatrix

setRow

public void setRow(int i,
                   AVector v)

Description copied from class: AMatrix

Sets a row in a matrix to the value specified by the given vector

Overrides:

setRow in class AMatrix

addAt

public void addAt(int i,
                  int j,
                  double d)

Description copied from interface: IMatrix

Adds a value at a specific position in the matrix, mutating the matrix Does not perform bounds checking - this in an unsafe operation

Specified by:

addAt in interface IMatrix

Overrides:

addAt in class AMatrix

addToArray

public void addToArray(double[] targetData,
                       int offset)

Description copied from interface: INDArray

Adds all the elements of this array to a double array at the specified offset, in row-major order

Specified by:

addToArray in interface INDArray

Overrides:

addToArray in class AMatrix

getColumns

public java.util.List<AVector> getColumns()

Description copied from interface: IMatrix

Gets a List of columns of a matrix. May return either copies or views, depending on the specific matrix type.

Specified by:

getColumns in interface IMatrix

Overrides:

getColumns in class AMatrix

Returns:

toSparseColumnMatrix

public SparseColumnMatrix toSparseColumnMatrix()

getRow

public AVector getRow(int i)

Description copied from interface: IMatrix

Returns a row of the matrix. May or may not be a view, depending on matrix type. Intended for the fastest possible read access of the row. This often means a view, but might not be (e.g. getRow on a Matrix33 returns a Vector3).

Specified by:

getRow in interface IMatrix

Specified by:

getRow in interface IFastRows

Overrides:

getRow in class AMatrix

getRows

public java.util.List<AVector> getRows()

Description copied from interface: IMatrix

Gets a List of rows of a matrix. May return either copies or views, depending on the specific matrix type.

Specified by:

getRows in interface IMatrix

Overrides:

getRows in class AMatrix

Returns:

getRowView

public AVector getRowView(int i)

Description copied from interface: IMatrix

Returns a row of the matrix as a vector view. May be used to modify the original matrix.

Specified by:

getRowView in interface IMatrix

Overrides:

getRowView in class AMatrix

isUpperTriangular

public boolean isUpperTriangular()

Description copied from class: AMatrix

Returns true if a matrix is upper triangular An upper triangular matrix is defined as having all elements equal to 0.0 where i > j

Overrides:

isUpperTriangular in class AMatrix

swapRows

public void swapRows(int i,
                     int j)

Description copied from class: AMatrix

Swaps two rows of the matrix in place This is an elementary row operation

Overrides:

swapRows in class AMatrix

replaceRow

public void replaceRow(int i,
                       AVector vec)

Description copied from class: AMatrix

Replaces a row in a matrix, adding the row to the internal structure of the matrix. Will throw UnsupportedOperationException if not possible for the given matrix type.

Overrides:

replaceRow in class AMatrix

add

public void add(AMatrix a)

Description copied from class: AMatrix

Adds another matrix to this matrix. Matrices must be the same size.

Overrides:

add in class AMatrix

copyRowTo

public void copyRowTo(int i,
                      double[] data,
                      int offset)

Description copied from class: AMatrix

Copies the elements in a selected row of this matrix to a double array

Overrides:

copyRowTo in class ASparseRCMatrix

Parameters:

i - The index of the selected row

data - Destination double[] array

offset - Offset into destination array

copyColumnTo

public void copyColumnTo(int col,
                         double[] targetData,
                         int offset)

Description copied from class: AMatrix

Copies the elements in a selected row of this matrix to a double array

Overrides:

copyColumnTo in class ASparseRCMatrix

targetData - Destination double[] array

getTransposeView

public SparseColumnMatrix getTransposeView()

Description copied from interface: INDArray

Returns a transposed view of the array. May throw UnsupportedOperationException if the array type does not support this capability

Specified by:

getTransposeView in interface INDArray

Specified by:

getTransposeView in interface IMatrix

Overrides:

getTransposeView in class AMatrix

Returns:

the transpose of this matrix, as a view

multiplyCopy

public AMatrix multiplyCopy(double a)

Description copied from interface: INDArray

Returns a copy of the array with all elements multiplied by a single constant value

Specified by:

multiplyCopy in interface INDArray

Overrides:

multiplyCopy in class AMatrix

Parameters:

a - A scalar factor

Returns:

A new array, with all element values scaled by the given factor

multiplyRow

public void multiplyRow(int i,
                        double value)

Description copied from class: AMatrix

Multiplies a row by a constant factor This is an elementary row operation

Overrides:

multiplyRow in class AMatrix

innerProduct

public AVector innerProduct(AVector a)

Description copied from interface: INDArray

Computes the inner product of this array with a vector

Specified by:

innerProduct in interface INDArray

Specified by:

innerProduct in interface IMatrix

Overrides:

innerProduct in class AMatrix

Parameters:

a - A vector

Returns:

A new array representing the inner product

rowDotProduct

public double rowDotProduct(int i,
                            AVector a)

Description copied from class: AMatrix

Returns the dot product of a specific row with a vector. Unsafe operation: performs no bounds checking

Overrides:

rowDotProduct in class AMatrix

Returns:

toDoubleArray

public double[] toDoubleArray()

Description copied from interface: INDArray

Copies the elements of this array to a new double[] array.

Specified by:

toDoubleArray in interface INDArray

Overrides:

toDoubleArray in class ASparseRCMatrix

innerProduct

public AMatrix innerProduct(AMatrix a)

Description copied from interface: IMatrix

Computes the inner product of this matrix with another. Equivalent to matrix x matrix multiplication

Specified by:

innerProduct in interface IMatrix

Overrides:

innerProduct in class AMatrix

Returns:

innerProduct

public SparseRowMatrix innerProduct(SparseColumnMatrix a)

Specialised inner product for sparse row matrix multiplied by sparse column matrix.

Parameters:

a -

Returns:

innerProduct

public SparseRowMatrix innerProduct(SparseRowMatrix a)

exactClone

public SparseRowMatrix exactClone()

Description copied from interface: INDArray

Creates a deep clone of an array, using the same class implementation as the original array

Specified by:

exactClone in interface INDArray

Specified by:

exactClone in class AMatrix

Returns:

A clone of the original array

clone

public AMatrix clone()

Description copied from interface: INDArray

Returns a clone of the array data, as a new array which will be fully mutable and may be of a different class to the original. Clone should attempt to return the most efficient possible array type. Clone should preserve sparsity property where possible, but this is not guaranteed.

Specified by:

clone in interface INDArray

Overrides:

clone in class AMatrix

sparse

public AMatrix sparse()

Description copied from interface: INDArray

Coerces this INDArray to a sparse format, without changing its element values. May return the same INDArray if already sparse. May also mutate the internal structure of the original NDArray, or create a view over parts of the original INDArray. You should take a defensive copy of the original NDArray if any of this concerns you. The returned sparse array may not be fully mutable in all elements.

Specified by:

sparse in interface INDArray

Overrides:

sparse in class ASparseRCMatrix

validate

public void validate()

Description copied from interface: INDArray

Validates the internal data structure of the INDArray. Throws an exception on failure. Failure indicates a serious bug and/or data corruption.

Specified by:

validate in interface INDArray

Overrides:

validate in class ASparseRCMatrix

equals

public boolean equals(AMatrix m)

Description copied from class: AMatrix

Returns true if this matrix is exactly equal to another matrix

Overrides:

equals in class AMatrix

equals

public boolean equals(IFastRows m)

Compares this matrix to another matrix implementing IFastColumns This can be much faster than the default equals

Parameters:

m -

Returns: