Class SquareMatrixSupport.ArrayRealSquareMatrix

java.lang.Object

org.apache.commons.statistics.inference.SquareMatrixSupport.ArrayRealSquareMatrix

All Implemented Interfaces:

SquareMatrixSupport.RealSquareMatrix

Enclosing class:

SquareMatrixSupport

private static class SquareMatrixSupport.ArrayRealSquareMatrix
extends Object
implements SquareMatrixSupport.RealSquareMatrix

Implementation of SquareMatrixSupport.RealSquareMatrix using a double[] array to store entries. Values are addressed using i*dim + j where dim is the square dimension.

Scaling is supported using the central element [m][m] where m = dimension/2. Scaling is only implemented post-multiplication to protect against overflow during repeat multiplication operations.

Note: The scaling is implemented to support computation of Kolmogorov's distribution as described in:

• Marsaglia, G., Tsang, W. W., & Wang, J. (2003). Evaluating Kolmogorov's Distribution. Journal of Statistical Software, 8(18), 1–4.

Since:

1.1

Field Summary

Fields

Modifier and Type	Field	Description
private final double[]	data	Entries of the matrix.
private final int	dim	Dimension.
private final int	exp	Matrix scale.
private static final double	SCALE_THRESHOLD	The scaling threshold.

Constructor Summary

Constructors

Constructor	Description
ArrayRealSquareMatrix(int dimension, double[] data, int scale)

Method Summary

Modifier and Type	Method	Description
private static void	checkExponent(int p)	Check the exponent is not negative.
int	dimension()	Gets the dimension for the rows and columns.
double	get(int i, int j)	Gets the value.
private SquareMatrixSupport.RealSquareMatrix	identity()	Creates the identity matrix I with the same dimension as this.
private static int	multiply(double[] a, int sa, double[] b, int sb, double[] col, double[] out)	Returns the result of postmultiplying a by b.
SquareMatrixSupport.RealSquareMatrix	power(int n)	Returns the result of multiplying this with itself n times.
int	scale()	Gets the scale of the matrix values.

Methods inherited from class Object

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

Field Details

SCALE_THRESHOLD

private static final double SCALE_THRESHOLD

The scaling threshold. Marsaglia used 1e140. This uses 2^400 ~ 2.58e120

See Also:

• Constant Field Values

dim

private final int dim

Dimension.

data

private final double[] data

Entries of the matrix.

exp

private final int exp

Matrix scale.

Constructor Details

ArrayRealSquareMatrix

ArrayRealSquareMatrix(int dimension,
 double[] data,
 int scale)

Parameters:

dimension - Matrix dimension.

data - Matrix data.

scale - Matrix scale.

Method Details

dimension

public int dimension()

Description copied from interface: SquareMatrixSupport.RealSquareMatrix

Gets the dimension for the rows and columns.

Specified by:

dimension in interface SquareMatrixSupport.RealSquareMatrix

Returns:

the dimension

scale

public int scale()

Description copied from interface: SquareMatrixSupport.RealSquareMatrix

Gets the scale of the matrix values. The true value is the value returned from SquareMatrixSupport.RealSquareMatrix.get(int, int) multiplied by 2^scale.

Specified by:

scale in interface SquareMatrixSupport.RealSquareMatrix

Returns:

the scale

get

public double get(int i,
 int j)

Description copied from interface: SquareMatrixSupport.RealSquareMatrix

Gets the value. This is a scaled value. The true value is the value returned multiplied by 2^scale.

Specified by:

get in interface SquareMatrixSupport.RealSquareMatrix

Parameters:

i - Row

j - Column

Returns:

the value

See Also:

• SquareMatrixSupport.RealSquareMatrix.scale()

power

public SquareMatrixSupport.RealSquareMatrix power(int n)

Description copied from interface: SquareMatrixSupport.RealSquareMatrix

Returns the result of multiplying this with itself n times.

Specified by:

power in interface SquareMatrixSupport.RealSquareMatrix

Parameters:

n - raise this to power n

Returns:

this^n

identity

private SquareMatrixSupport.RealSquareMatrix identity()

Creates the identity matrix I with the same dimension as this.

Returns:

I

multiply

private static int multiply(double[] a,
 int sa,
 double[] b,
 int sb,
 double[] col,
 double[] out)

Returns the result of postmultiplying a by b. It is expected the scale of the result will be the sum of the scale of the arguments; this may be adjusted by the scale power if the result is scaled by a power of two for overflow protection.

Parameters:

a - Matrix.

sa - Scale of matrix a.

b - Matrix to postmultiply by.

sb - Scale of matrix b.

col - Working array for a column of the matrix.

out - Output a * b

Returns:

Scale of a * b

checkExponent

private static void checkExponent(int p)

Check the exponent is not negative.

Parameters:

p - Exponent.

Throws:

IllegalArgumentException - if the exponent is negative