org.ojalgo.matrix.jama

Class JamaSingularValue

java.lang.Object

org.ojalgo.matrix.jama.JamaSingularValue

All Implemented Interfaces:

MatrixDecomposition<Double>, SingularValue<Double>, DecompositionTask<Double>, InverterTask<Double>, SolverTask<Double>

public final class JamaSingularValue
extends Object
implements SingularValue<Double>

This class adapts JAMA's SingularValueDecomposition to ojAlgo's SingularValue interface. speed: 52.641s

Author:

apete

Nested Class Summary

Nested classes/interfaces inherited from interface org.ojalgo.matrix.decomposition.task.InverterTask

InverterTask.Factory<N extends Number>

Nested classes/interfaces inherited from interface org.ojalgo.matrix.decomposition.task.SolverTask

SolverTask.Factory<N extends Number>

Field Summary

Fields inherited from interface org.ojalgo.matrix.decomposition.task.InverterTask

BIG, COMPLEX, PRIMITIVE

Fields inherited from interface org.ojalgo.matrix.decomposition.task.SolverTask

BIG, COMPLEX, PRIMITIVE

Constructor Summary

Constructors

Constructor and Description
JamaSingularValue() Not recommended to use this constructor directly.

Method Summary

Modifier and Type	Method and Description
boolean	compute(Access2D<?> aStore)
boolean	compute(Access2D<?> matrix, boolean singularValuesOnly, boolean fullSize)
boolean	equals(MatrixDecomposition<Double> other, NumberContext context)
boolean	equals(MatrixStore<Double> aStore, NumberContext context)
double	getCondition() The condition number.
JamaMatrix	getD()
double	getFrobeniusNorm() Sometimes also called the Schatten 2-norm or Hilbert-Schmidt norm.
JamaMatrix	getInverse() The output must be a "right inverse" and a "generalised inverse".
MatrixStore<Double>	getInverse(DecompositionStore<Double> preallocated) Makes no use of preallocated at all.
double	getKyFanNorm(int k) Ky Fan k-norm.
double	getOperatorNorm()
JamaMatrix	getQ1() If [A] is m-by-n and its rank is r, then: The first r columns of [Q1] span the column space, range or image of [A]. The last m-r columns of [Q1] span the left nullspace or cokernel of [A]. Calculating the QR decomposition of [A] is a faster alternative.
JamaMatrix	getQ2() If [A] is m-by-n and its rank is r, then: The first r columns of [Q2] span the row space or coimage of [A]. The last n-r columns of [Q2] span the nullspace or kernel of [A]. Calculating the QR decomposition of [A]T is a faster alternative.
int	getRank() Effective numerical matrix rank.
Array1D<Double>	getSingularValues()
double	getTraceNorm()
MatrixStore<Double>	invert(MatrixStore<Double> original) The output must be a "right inverse" and a "generalised inverse".
MatrixStore<Double>	invert(MatrixStore<Double> original, DecompositionStore<Double> preallocated) Implementiong this method is optional.
boolean	isAspectRatioNormal()
boolean	isComputed()
boolean	isFullSize()
boolean	isOrdered()
boolean	isSolvable()
DecompositionStore<Double>	preallocate(Access2D<Double> template) Implementiong this method is optional.
DecompositionStore<Double>	preallocate(Access2D<Double> templateBody, Access2D<Double> templateRHS) Implementiong this method is optional.
MatrixStore<Double>	reconstruct()
void	reset() Delete computed results, and resets attributes to default values
JamaMatrix	solve(Access2D<Double> rhs) Internally this implementation uses the pseudoinverse that is recreated with every call.
MatrixStore<Double>	solve(Access2D<Double> body, Access2D<Double> rhs) [A][X]=[B] or [this][return]=[aRHS]
MatrixStore<Double>	solve(Access2D<Double> body, Access2D<Double> rhs, DecompositionStore<Double> preallocated) Implementiong this method is optional.
JamaMatrix	solve(Access2D<Double> rhs, DecompositionStore<Double> preallocated) Makes no use of preallocated at all.
String	toString()

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.ojalgo.matrix.decomposition.MatrixDecomposition

compute, equals, getInverse, preallocate, solve

Methods inherited from interface org.ojalgo.matrix.decomposition.task.InverterTask

invert, invert, preallocate

Methods inherited from interface org.ojalgo.matrix.decomposition.task.SolverTask

solve, solve

Constructor Detail

JamaSingularValue

public JamaSingularValue()

Not recommended to use this constructor directly. Consider using the static factory method SingularValueDecomposition.makeJama() instead.

Method Detail

compute

public boolean compute(Access2D<?> matrix,
                       boolean singularValuesOnly,
                       boolean fullSize)

Specified by:

compute in interface SingularValue<Double>

Parameters:

matrix - A matrix to decompose

singularValuesOnly - No need to calculate eigenvectors

fullSize - TODO

Returns:

true/false if the computation succeeded or not

equals

public boolean equals(MatrixStore<Double> aStore,
                      NumberContext context)

Specified by:

equals in interface MatrixDecomposition<Double>

getCondition

public double getCondition()

Description copied from interface: SingularValue

The condition number.

Specified by:

getCondition in interface SingularValue<Double>

Returns:

The largest singular value divided by the smallest singular value.

getD

public JamaMatrix getD()

Specified by:

getD in interface SingularValue<Double>

Returns:

The diagonal matrix of singular values.

getFrobeniusNorm

public double getFrobeniusNorm()

Description copied from interface: SingularValue

Sometimes also called the Schatten 2-norm or Hilbert-Schmidt norm.

Specified by:

getFrobeniusNorm in interface SingularValue<Double>

Returns:

The square root of the sum of squares of the singular values.

getInverse

public JamaMatrix getInverse()

Description copied from interface: MatrixDecomposition

The output must be a "right inverse" and a "generalised inverse".

Specified by:

getInverse in interface MatrixDecomposition<Double>

See Also:

BasicMatrix.invert()

getKyFanNorm

public double getKyFanNorm(int k)

Description copied from interface: SingularValue

Ky Fan k-norm.

The first Ky Fan k-norm is the operator norm (the largest singular value), and the last is called the trace norm (the sum of all singular values).

Specified by:

getKyFanNorm in interface SingularValue<Double>

Parameters:

k - The number of singular values to add up.

Returns:

The sum of the k largest singular values.

getOperatorNorm

public double getOperatorNorm()

Specified by:

getOperatorNorm in interface SingularValue<Double>

Returns:

2-norm

getQ1

public JamaMatrix getQ1()

Description copied from interface: SingularValue

If [A] is m-by-n and its rank is r, then:

• The first r columns of [Q1] span the column space, range or image of [A].
• The last m-r columns of [Q1] span the left nullspace or cokernel of [A].

Calculating the QR decomposition of [A] is a faster alternative.

Specified by:

getQ1 in interface SingularValue<Double>

getQ2

public JamaMatrix getQ2()

Description copied from interface: SingularValue

If [A] is m-by-n and its rank is r, then:

• The first r columns of [Q2] span the row space or coimage of [A].
• The last n-r columns of [Q2] span the nullspace or kernel of [A].

Calculating the QR decomposition of [A]^T is a faster alternative.

Specified by:

getQ2 in interface SingularValue<Double>

getRank

public int getRank()

Description copied from interface: SingularValue

Effective numerical matrix rank.

Specified by:

getRank in interface SingularValue<Double>

Returns:

The number of nonnegligible singular values.

getSingularValues

public Array1D<Double> getSingularValues()

Specified by:

getSingularValues in interface SingularValue<Double>

Returns:

The singular values ordered in descending order.

getTraceNorm

public double getTraceNorm()

Specified by:

getTraceNorm in interface SingularValue<Double>

isAspectRatioNormal

public boolean isAspectRatioNormal()

isComputed

public boolean isComputed()

Specified by:

isComputed in interface MatrixDecomposition<Double>

Returns:

true if computation has been attemped; false if not.

See Also:

MatrixDecomposition.compute(Access2D), MatrixDecomposition.isSolvable()

isFullSize

public boolean isFullSize()

Specified by:

isFullSize in interface MatrixDecomposition<Double>

Returns:

True if the implementation generates a full sized decomposition.

isOrdered

public boolean isOrdered()

Specified by:

isOrdered in interface SingularValue<Double>

isSolvable

public boolean isSolvable()

Specified by:

isSolvable in interface MatrixDecomposition<Double>

Returns:

true if it is ok to call MatrixDecomposition.solve(Access2D) (computation was successful); false if not

See Also:

MatrixDecomposition.solve(Access2D), MatrixDecomposition.isComputed()

reconstruct

public MatrixStore<Double> reconstruct()

Specified by:

reconstruct in interface MatrixDecomposition<Double>

reset

public void reset()

Description copied from interface: MatrixDecomposition

Delete computed results, and resets attributes to default values

Specified by:

reset in interface MatrixDecomposition<Double>

solve

public JamaMatrix solve(Access2D<Double> rhs)

Internally this implementation uses the pseudoinverse that is recreated with every call.

Specified by:

solve in interface MatrixDecomposition<Double>

toString

public String toString()

Overrides:

toString in class Object

compute

public final boolean compute(Access2D<?> aStore)

Specified by:

compute in interface MatrixDecomposition<Double>

Parameters:

aStore - A matrix to decompose

Returns:

true if the computation suceeded; false if not

equals

public final boolean equals(MatrixDecomposition<Double> other,
                            NumberContext context)

Specified by:

equals in interface MatrixDecomposition<Double>

getInverse

public final MatrixStore<Double> getInverse(DecompositionStore<Double> preallocated)

Makes no use of preallocated at all. Simply delegates to getInverse().

Specified by:

getInverse in interface MatrixDecomposition<Double>

Parameters:

preallocated - Preallocated memory for the results, possibly some intermediate results. You must assume this is modified, but you cannot assume it will contain the full/final/correct solution.

Returns:

The inverse

See Also:

MatrixDecomposition.getInverse(org.ojalgo.matrix.decomposition.DecompositionStore)

invert

public final MatrixStore<Double> invert(MatrixStore<Double> original)

Description copied from interface: InverterTask

The output must be a "right inverse" and a "generalised inverse".

Specified by:

invert in interface InverterTask<Double>

See Also:

BasicMatrix.invert()

invert

public final MatrixStore<Double> invert(MatrixStore<Double> original,
                                        DecompositionStore<Double> preallocated)

Description copied from interface: InverterTask

Implementiong this method is optional.

Exactly how a specific implementation makes use of preallocated is not specified by this interface. It must be documented for each implementation.

Should produce the same results as calling #getInverse().

Specified by:

invert in interface InverterTask<Double>

preallocated - Preallocated memory for the results, possibly some intermediate results. You must assume this is modified, but you cannot assume it will contain the full/final/correct solution.

Returns:

The inverse

preallocate

public final DecompositionStore<Double> preallocate(Access2D<Double> template)

Description copied from interface: InverterTask

Implementiong this method is optional.

Will create a DecompositionStore instance suitable for use with #solve(Access2D, DecompositionStore). When solving an equation system [A][X]=[B] ([mxn][nxb]=[mxb]) the preallocated memory/matrix will typically be either mxb or nxb (if A is square then there is no doubt).

Specified by:

preallocate in interface InverterTask<Double>

Returns:

preallocate

public final DecompositionStore<Double> preallocate(Access2D<Double> templateBody,
                                                    Access2D<Double> templateRHS)

Description copied from interface: MatrixDecomposition

Implementiong this method is optional.

Will create a DecompositionStore instance suitable for use with MatrixDecomposition.solve(Access2D, DecompositionStore). When solving an equation system [A][X]=[B] ([mxn][nxb]=[mxb]) the preallocated memory/matrix will typically be either mxb or nxb (if A is square then there is no doubt).

Specified by:

preallocate in interface MatrixDecomposition<Double>

Specified by:

preallocate in interface SolverTask<Double>

Returns:

solve

public final MatrixStore<Double> solve(Access2D<Double> body,
                                       Access2D<Double> rhs)

Description copied from interface: SolverTask

[A][X]=[B] or [this][return]=[aRHS]

Specified by:

solve in interface SolverTask<Double>

solve

public final MatrixStore<Double> solve(Access2D<Double> body,
                                       Access2D<Double> rhs,
                                       DecompositionStore<Double> preallocated)

Description copied from interface: SolverTask

Implementiong this method is optional.

Exactly how a specific implementation makes use of preallocated is not specified by this interface. It must be documented for each implementation.

Should produce the same results as calling #solve(Access2D).

Specified by:

solve in interface SolverTask<Double>

rhs - The Right Hand Side, wont be modfied

preallocated - Preallocated memory for the results, possibly some intermediate results. You must assume this is modified, but you cannot assume it will contain the full/final/correct solution.

Returns:

The solution

solve

public final JamaMatrix solve(Access2D<Double> rhs,
                              DecompositionStore<Double> preallocated)

Makes no use of preallocated at all. Simply delegates to solve(Access2D).

Specified by:

solve in interface MatrixDecomposition<Double>

Parameters:

rhs - The Right Hand Side, wont be modfied

preallocated - Preallocated memory for the results, possibly some intermediate results. You must assume this is modified, but you cannot assume it will contain the full/final/correct solution.

Returns:

The solution

See Also:

MatrixDecomposition.solve(Access2D, org.ojalgo.matrix.decomposition.DecompositionStore)