jsat.testing.goodnessoffit

Class KSTest

java.lang.Object

jsat.testing.goodnessoffit.KSTest

public class KSTest
extends Object

Author:

Edward Raff

Constructor Summary

Constructors

Constructor and Description
KSTest(Vec v) Creates a new statistical test for testing.

Method Summary

Modifier and Type	Method and Description
protected double	dCalc(ContinuousDistribution cd) Calculates the D statistic for comparison against a continous distribution
protected double	dCaldO(Vec o) Calculates the D statistic for comparison against another data set
void	setBaseData(Vec v) Change the original sample to v
double	testData(Vec data) Returns the p-value for the 2 sample KS Test against the given data set data.
double	testDist(ContinuousDistribution cd) Returns the p-value for the KS Test against the given distribution cd.

Methods inherited from class java.lang.Object

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

Constructor Detail

KSTest

public KSTest(Vec v)

Creates a new statistical test for testing. The 1 sample test, with v being the 1 sample. The 1 sample test compare the data to a given distribution, and see if it does not belong to the given distribution. The 2 sample test is designed to tell if the data is not from the same population.

Parameters:

v - the date to be one of the samples

Method Detail

setBaseData

public void setBaseData(Vec v)

Change the original sample to v

Parameters:

v - the new original sample.

dCalc

protected double dCalc(ContinuousDistribution cd)

Calculates the D statistic for comparison against a continous distribution

Parameters:

cd - the distribution to compare against

Returns:

the max difference between the empirical CDF and the 'true' CDF of the given distribution

dCaldO

protected double dCaldO(Vec o)

Calculates the D statistic for comparison against another data set

Parameters:

o - the other data set

Returns:

the max difrence in empirical CDF of the distributions.

testDist

public double testDist(ContinuousDistribution cd)

Returns the p-value for the KS Test against the given distribution cd.
The null hypothesis of this test is that the given data set belongs to the given distribution.
The alternative hypothesis is that the data set does not belong to the given distribution.

Parameters:

cd - the distribution to compare against

Returns:

the p-value of the test against this distribution

testData

public double testData(Vec data)

Returns the p-value for the 2 sample KS Test against the given data set data.
The null hypothesis of this test is that the given data set is from the same population as data
The alternative hypothesis is that the data set does not belong to the same population as data

Parameters:

data - the other distribution to compare against

Returns:

the p-value of the test against this data set