Package ubic.basecode.math

Class KruskalWallis

java.lang.Object

ubic.basecode.math.KruskalWallis

public class KruskalWallis
extends Object

Perform a Kruskal-Wallis test.

Author:

paul

Constructor Summary

Constructors

Constructor	Description
KruskalWallis()

Method Summary

Modifier and Type	Method	Description
static int	dof(cern.colt.list.DoubleArrayList scores, cern.colt.list.IntArrayList groupings)
static double	kwStatistic(cern.colt.list.DoubleArrayList scores, cern.colt.list.IntArrayList groupings)
static double	test(cern.colt.list.DoubleArrayList scores, cern.colt.list.IntArrayList groupings)	Perform a Kruskal-Wallis one-way ANOVA.

Methods inherited from class java.lang.Object

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

Constructor Detail

KruskalWallis

public KruskalWallis()

Method Detail

dof

public static int dof(cern.colt.list.DoubleArrayList scores,
                      cern.colt.list.IntArrayList groupings)

Parameters:

scores -

groupings -

Returns:

number of degrees of freedom (number of groups - 1)

kwStatistic

public static double kwStatistic(cern.colt.list.DoubleArrayList scores,
                                 cern.colt.list.IntArrayList groupings)

Parameters:

scores -

groupings -

Returns:

statistic; chi-squared with numgroups - 1 dof under the null.

test

public static double test(cern.colt.list.DoubleArrayList scores,
                          cern.colt.list.IntArrayList groupings)

Perform a Kruskal-Wallis one-way ANOVA.

Implementation note: Does not make special corrections for ties, though ties are given the averaged ranks. Completely bare-bones. Missing values are not tolerated well.

Parameters:

scores -

groupings - integer indicators of which values are in which groups. The actual values don't matter.

Returns:

p-values based on the chi-squared statistic.