org.apache.mahout.math

Class CholeskyDecomposition

java.lang.Object

org.apache.mahout.math.CholeskyDecomposition

public class CholeskyDecomposition
extends Object

Cholesky decomposition shamelessly ported from JAMA.

A Cholesky decomposition of a semi-positive definite matrix A is a lower triangular matrix L such that L L^* = A. If A is full rank, L is unique. If A is real, then it must be symmetric and R will also be real.

Constructor Summary

Constructors

Constructor and Description
CholeskyDecomposition(Matrix a)
CholeskyDecomposition(Matrix a, boolean pivot)

Method Summary

Methods

Modifier and Type	Method and Description
int[]	getInversePivot()
Matrix	getL()
PivotedMatrix	getPermutedL()
int[]	getPivot()
boolean	isPositiveDefinite()
Matrix	solveLeft(Matrix z) Compute inv(L) * z efficiently.
Matrix	solveRight(Matrix z) Compute z * inv(L') efficiently

Methods inherited from class java.lang.Object

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

Constructor Detail

CholeskyDecomposition

public CholeskyDecomposition(Matrix a)

CholeskyDecomposition

public CholeskyDecomposition(Matrix a,
                     boolean pivot)

Method Detail

isPositiveDefinite

public boolean isPositiveDefinite()

getL

public Matrix getL()

getPermutedL

public PivotedMatrix getPermutedL()

getPivot

public int[] getPivot()

Returns:

Returns the permutation of rows and columns that was applied to L

getInversePivot

public int[] getInversePivot()

solveLeft

public Matrix solveLeft(Matrix z)

Compute inv(L) * z efficiently.

Parameters:

z -

solveRight

public Matrix solveRight(Matrix z)

Compute z * inv(L') efficiently