Package uk.ac.ebi.utils.statistics

Class FisherExact

java.lang.Object
uk.ac.ebi.utils.statistics.FisherExact
public class FisherExact extends Object
Class for Fisher Exact test.

This was copied and adapted from here, (commit 71b68de, 2023/07/17), since the original class is part of a bigger project concerning a different specific application domain.

TODO: It would be better to have a cached object that depends on maxSize as cache key, ie, when a requested maxSize is < cached maxSize, the existing instance is returned, else a new instance is created that replaces the old one. This also requires that such instance has an expiration time, a cache that evicts automatically and an ability to initialise a FisherExact with a bigger size from one with a smaller one. At that point, we would prefer static wrapper methods for methods like getP(int, int, int, int).

Original comments

This does a Fisher Exact test. The Fisher's Exact test procedure calculates an exact probability value for the relationship between two dichotomous variables, as found in a two by two crosstable. The program calculates the difference between the data observed and the data expected, considering the given marginal and the assumptions of the model of independence. It works in exactly the same way as the Chi-square test for independence; however, the Chi-square gives only an estimate of the true probability value, an estimate which might not be very accurate if the marginal is very uneven or if there is a small value (less than five) in one of the cells. It uses an array of factorials initialized at the beginning to provide speed. There could be better ways to do this.
Version:
$Id: FisherExact.java,v 1
Author:
Ed Buckler

  • Constructor Summary

    Constructors
    Constructor
    Description
    FisherExact(int maxSize)
    constructor for FisherExact table

  • Method Summary

    Modifier and Type
    Method
    Description
    final double
    getCumlativeP(int a, int b, int c, int d)
    Calculates the one-tail P-value for the Fisher Exact test.
    final double
    getLeftTailedP(int a, int b, int c, int d)
    Calculates the left-tail P-value for the Fisher Exact test.
    static double
    getOddsRatio(double a, double b, double c, double d)
    Measure of how far from independence the 2x2 table is. 1= independent.
    static double[]
    getOddsRatioAnd95thConfidenceInterval(double a, double b, double c, double d)
    Returns OR, lower, upper
    final double
    getP(int a, int b, int c, int d)
    calculates the P-value for this specific state
    final double
    getRightTailedP(int a, int b, int c, int d)
    Calculates the right-tail P-value for the Fisher Exact test.
    final double
    getTwoTailedP(int a, int b, int c, int d)
    Calculates the two-tailed P-value for the Fisher Exact test.

    Methods inherited from class java.lang.Object

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

  • Constructor Details

    • FisherExact

      public FisherExact(int maxSize)
      constructor for FisherExact table
      Parameters:
      maxSize - is the maximum sum that will be encountered by the table (a+b+c+d)

  • Method Details

    • getP

      public final double getP(int a, int b, int c, int d)
      calculates the P-value for this specific state
      Parameters:
      a - a, b, c, d are the four cells in a 2x2 matrix
      b -
      c -
      d -
      Returns:
      the P-value

    • getCumlativeP

      public final double getCumlativeP(int a, int b, int c, int d)
      Calculates the one-tail P-value for the Fisher Exact test. Determines whether to calculate the right- or left- tail, thereby always returning the smallest p-value. a, b, c, d are the four cells in a 2x2 matrix
      Returns:
      one-tailed P-value (right or left, whichever is smallest)

    • getRightTailedP

      public final double getRightTailedP(int a, int b, int c, int d)
      Calculates the right-tail P-value for the Fisher Exact test. a, b, c, d are the four cells in a 2x2 matrix
      Returns:
      one-tailed P-value (right-tail)

    • getLeftTailedP

      public final double getLeftTailedP(int a, int b, int c, int d)
      Calculates the left-tail P-value for the Fisher Exact test. a, b, c, d are the four cells in a 2x2 matrix
      Returns:
      one-tailed P-value (left-tail)

    • getTwoTailedP

      public final double getTwoTailedP(int a, int b, int c, int d)
      Calculates the two-tailed P-value for the Fisher Exact test. In order for a table under consideration to have its p-value included in the final result, it must have a p-value less than the original table's P-value, i.e. Fisher's exact test computes the probability, given the observed marginal frequencies, of obtaining exactly the frequencies observed and any configuration more extreme. By "more extreme," we mean any configuration (given observed marginals) with a smaller probability of occurrence in the same direction (one-tailed) or in both directions (two-tailed). a, b, c, d are the four cells in a 2x2 matrix
      Returns:
      two-tailed P-value or NaN if the table sum exceeds the maxSize

    • getOddsRatioAnd95thConfidenceInterval

      public static double[] getOddsRatioAnd95thConfidenceInterval(double a, double b, double c, double d)
      Returns OR, lower, upper

    • getOddsRatio

      public static double getOddsRatio(double a, double b, double c, double d)
      Measure of how far from independence the 2x2 table is. 1= independent. Ratio of ratios. Will return Infinity or 0 if cells are zero.