pal.statistics
Class ContigencyTable

java.lang.Object
  pal.statistics.ContigencyTable

public class ContigencyTable
extends java.lang.Object

Class for permuting contigency tables and determining the likelihood of the table. If determining of the probability of a 2x2 table use FisherExact as it is much faster.

Version:

$Id: ContigencyTable.java,v 1

Author:

Ed Buckler

Constructor Summary

ContigencyTable(int maxSize)
constructor for Contigency table

Method Summary

double	calcContigencyChiSquare(int permutations) This calculates the probability in the normal approach, using the Chi Square as the test statistic.
double	calcMonteCarloExactTest(int permutations) This calculates the probability in the normal permutation approach, using the method described by Weir, B.
double	calcRapidContigencyChiSquare(int maxPermutations) This calculates the probability in a rapid approach, using the Chi Square as the test statistic.
double	calcRapidMonteCarloExactTest(int maxPermutations) This calculates the probability in the rapid permutational approach, using the method described by Weir, B.
void	setMatrix(int[][] tcontig) sets the data for the contigency table, must be set before other methods are called.

Methods inherited from class java.lang.Object

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

Constructor Detail

ContigencyTable

public ContigencyTable(int maxSize)

constructor for Contigency table

Parameters:

maxSize - is the maximum sum that will be encountered by contigency table

Method Detail

setMatrix

public void setMatrix(int[][] tcontig)

sets the data for the contigency table, must be set before other methods are called. If tcontig has a greater count than maxSize, then the contig is set to null

Parameters:

contig - is the array of integers with observed states

calcRapidContigencyChiSquare

public double calcRapidContigencyChiSquare(int maxPermutations)

This calculates the probability in a rapid approach, using the Chi Square as the test statistic. It runs for maxPermutations permutations unless it find 10 values that beat the observed, and then it stops and calculates the p-value. This slighly biases the P-values but makes it much more rapid.

Parameters:

maxPermutations - Number of permutations used to calculate the probability

Returns:

P-value (NaN is returned if bad contig was set)

calcContigencyChiSquare

public double calcContigencyChiSquare(int permutations)

This calculates the probability in the normal approach, using the Chi Square as the test statistic.

Parameters:

permutations - Number of permutations used to calculate the probability

Returns:

P-value (NaN is returned if bad contig was set)

calcRapidMonteCarloExactTest

public double calcRapidMonteCarloExactTest(int maxPermutations)

This calculates the probability in the rapid permutational approach, using the method described by Weir, B. S. (1996) Genetic Data Analysis II (Sinauer, Sunderland, MA) It runs for 1000 permutations unless it find 10 values that beat the observed, and then it stops and calculates the p-value. This slighly biases the P-values but makes it much more rapid.

Parameters:

reps - is the number of permutations used to the probability

Returns:

P-value (NaN is returned if bad contig was set)

calcMonteCarloExactTest

public double calcMonteCarloExactTest(int permutations)

This calculates the probability in the normal permutation approach, using the method described by Weir, B. S. (1996) Genetic Data Analysis II (Sinauer, Sunderland, MA).

Parameters:

permutations - Number of permutations used to the probability

Returns:

P-value (NaN is returned if bad contig was set)