Package org.apache.commons.statistics.inference

Class Hypergeom

java.lang.Object

org.apache.commons.statistics.inference.Hypergeom

class Hypergeom
extends java.lang.Object

Provide a wrapper around the HypergeometricDistribution that caches all probability mass values.

This class extracts the logic from the HypergeometricDistribution implementation used for the cumulative probability functions. It allows fast computation of the CDF and SF for the entire supported domain.

Since:

1.1

Field Summary

Fields

Modifier and Type	Field	Description
private static double	HALF	1/2.
private int	lowerBound	The lower bound of the support (inclusive).
private int	m	Cached midpoint, m, of the CDF/SF.
private int	m1	Lower mode.
private int	m2	Upper mode.
private double	midCDF	Cached CDF of the midpoint.
private double[]	prob	Cached probability values.
private int	upperBound	The upper bound of the support (inclusive).

Constructor Summary

Constructors

Constructor	Description
Hypergeom(int populationSize, int numberOfSuccesses, int sampleSize)

Method Summary

Modifier and Type	Method	Description
(package private) double	cdf(int x)	Compute the cumulative distribution function (CDF) at the specified value.
(package private) int	getLowerMode()	Get the lower mode of the distribution.
(package private) int	getSupportLowerBound()	Get the lower bound of the support.
(package private) int	getSupportUpperBound()	Get the upper bound of the support.
(package private) int	getUpperMode()	Get the upper mode of the distribution.
private double	innerCumulativeProbability(int x0, int x1)	For this distribution, X, this method returns P(x0 <= X <= x1).
(package private) double	pmf(int x)	Compute the probability mass function (PMF) at the specified value.
(package private) double	sf(int x)	Compute the survival function (SF) at the specified value.

Methods inherited from class java.lang.Object

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

Field Detail

HALF

private static final double HALF

1/2.

See Also:

Constant Field Values

lowerBound

private final int lowerBound

The lower bound of the support (inclusive).

upperBound

private final int upperBound

The upper bound of the support (inclusive).

prob

private final double[] prob

Cached probability values. This holds values from x=0 even though the supported lower bound may be above x=0. This allows x to be used as an index without offsetting using the lower bound.

m

private final int m

Cached midpoint, m, of the CDF/SF. This is not the true median. It is the value where the CDF is closest to 0.5; as such the CDF(m) may be below 0.5 if the next value CDF(m+1) is further from 0.5. Used for the cumulative probability functions.

midCDF

private final double midCDF

Cached CDF of the midpoint. Used for the cumulative probability functions.

m1

private final int m1

Lower mode.

m2

private final int m2

Upper mode.

Constructor Detail

Hypergeom

Hypergeom(int populationSize,
          int numberOfSuccesses,
          int sampleSize)

Parameters:

populationSize - Population size.

numberOfSuccesses - Number of successes in the population.

sampleSize - Sample size.

Method Detail

getSupportLowerBound

int getSupportLowerBound()

Get the lower bound of the support.

Returns:

lower bound

getSupportUpperBound

int getSupportUpperBound()

Get the upper bound of the support.

Returns:

upper bound

getLowerMode

int getLowerMode()

Get the lower mode of the distribution.

Returns:

lower mode

getUpperMode

int getUpperMode()

Get the upper mode of the distribution.

Returns:

upper mode

pmf

double pmf(int x)

Compute the probability mass function (PMF) at the specified value.

Parameters:

x - Value.

Returns:

P(X = x)

Throws:

java.lang.IndexOutOfBoundsException - if the value x is not in the supported domain.

cdf

double cdf(int x)

Compute the cumulative distribution function (CDF) at the specified value.

Parameters:

x - Value.

Returns:

P(X <= x)

sf

double sf(int x)

Compute the survival function (SF) at the specified value. This is the complementary cumulative distribution function.

Parameters:

x - Value.

Returns:

P(X > x)

innerCumulativeProbability

private double innerCumulativeProbability(int x0,
                                          int x1)

For this distribution, X, this method returns P(x0 <= X <= x1). This probability is computed by summing the point probabilities for the values x0, x0 + dx, x0 + 2 * dx, ..., x1; the direction dx is determined using a comparison of the input bounds. This should be called by using x0 as the domain limit and x1 as the internal value. This will result in a sum of increasingly larger magnitudes.

Parameters:

x0 - Inclusive domain bound.

x1 - Inclusive internal bound.

Returns:

P(x0 <= X <= x1).