/***********************************************************************
This file is part of JavaNPST, a Java library of NonParametric
Statistical Tests.
Copyright (C) 2011
J. Derrac (jderrac@decsai.ugr.es)
S. Garcia (sglopez@ujaen.es)
F. Herrera (herrera@decsai.ugr.es)
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see http://www.gnu.org/licenses/
**********************************************************************/
package javanpst.distributions.tests;
import java.io.IOException;
import javax.xml.parsers.ParserConfigurationException;
import javax.xml.parsers.SAXParser;
import javax.xml.parsers.SAXParserFactory;
import org.xml.sax.SAXException;
import javanpst.data.distributions.incompleteTable.Incomplete2KeyTable;
import javanpst.data.readers.distributionReaders.WilcoxonReaderFormat;
import javanpst.distributions.DistributionDefinitions;
import javanpst.distributions.common.continuous.NormalDistribution;
import javanpst.utils.Formats;
/**
* An implementation of the Wilcoxon Signed-ranks distribution.
*
* Uses a Singleton pattern to ensure only an instance of the class is allowed.
*
* @author Joaqu�n Derrac Rus (jderrac@decsai.ugr.es)
* @version 1.0
*/
public class WilcoxonDistribution extends TestDistribution{
/**
* Table of the distribution
*/
private static Incomplete2KeyTable table;
/**
* Normal distribution for adjusting
*/
private static NormalDistribution normal;
/**
* Private default builder.
*
* Initializes distribution tables
*/
private WilcoxonDistribution(){
initialize();
}//end-method
/**
* Singleton pattern access to the distribution
*
* @return the instance of the distribution
*/
public static WilcoxonDistribution getInstance(){
return SingletonHolder.distribution;
}//end-method
/**
* Singleton holder for the only instance of the distribution
*
* SingletonHolder is loaded on the first execution of getInstance()
* or the first access to SingletonHolder.distribution, not before.
*/
private static class SingletonHolder {
public static final WilcoxonDistribution distribution = new WilcoxonDistribution();
}//end-method
/**
* Generate distribution tables
*/
protected void generateTables(){
table=new Incomplete2KeyTable(505,51);
loadFiles();
}//end-method
/**
* Load data files to fill distribution tables
*/
protected void loadFiles(){
try{
SAXParserFactory spf=SAXParserFactory.newInstance();
SAXParser sp = spf.newSAXParser();
sp.parse(this.getClass().getResourceAsStream("/tables/Wilcoxon/WilcoxonTable.xml"), new WilcoxonReaderFormat());
}catch(ParserConfigurationException e){
System.err.println("Parser error");
}catch(SAXException e2){
System.err.println("SAX error: " + e2.getStackTrace());
} catch (IOException e3) {
System.err.println("Input/Output error: " + e3.getMessage() );
}
}//end-method
/**
* Gets the table of the distribution
*
* @return Reference to the table
*/
public static Incomplete2KeyTable getTable(){
return table;
}//end-method
/**
* Computes exact p-value of the Wilcoxon distribution
*
* @param N n parameter
* @param R rank associated
* @return p-value computed
*/
public double computeExactProbability(int N,double R){
double pValue,value1,value2;
if(N>50){
return DistributionDefinitions.UNDEFINED;
}
try{
if(Formats.isInt(R)){
pValue=table.get((int)R, N);
if(pValue==DistributionDefinitions.UNDEFINED){
pValue=DistributionDefinitions.ALL;
}
}else{
value1=table.get((int)Math.ceil(R), N);
if(value1==DistributionDefinitions.UNDEFINED){
value1=DistributionDefinitions.ALL;
}
value2=table.get((int)Math.floor(R), N);
if(value2==DistributionDefinitions.UNDEFINED){
value2=DistributionDefinitions.ALL;
}
pValue=(value1+value2)/2.0;
}
}
catch(ArrayIndexOutOfBoundsException e){
return DistributionDefinitions.ALL;
}
return pValue;
}//end-method
/**
* Computes asymptotic double tailed p-value of the Wilcoxon distribution
*
* A ties correction of variance is incorporated. The array ties
* must be in the form {2,3,2,3} where each integer denotes the occurrence of a tie
* and the number of elements involved.
*
* if no ties, a void array shoud be used (i.e. ties=new int [0];)
*
* @param N n parameter
* @param R rank associated
* @param ties weight of ties
* @return p-value computed
*/
public double computeAsymptoticDoubleTailProbability(int N,double R,double ties){
double numerator1,numerator2,denominator;
double z;
double leftTail, rightTail, doubleTail;
normal=new NormalDistribution();
numerator1=R-0.5-(N*(N+1.0)/4.0);
numerator2=R+0.5-(N*(N+1.0)/4.0);
denominator=Math.sqrt(N*(N+1.0)*((2*N)+1.0)/24.0);
denominator-=ties/48.0;
z=numerator1/denominator;
rightTail=1.0-normal.getTipifiedProbability(z, false);
z=numerator2/denominator;
leftTail=normal.getTipifiedProbability(z, false);
doubleTail=Math.min(leftTail, rightTail)*2.0;
doubleTail=Math.min(doubleTail,1.0);
return doubleTail;
}//end-method
/**
* Computes asymptotic right tailed p-value of the Wilcoxon distribution
*
* A ties correction of variance is incorporated. The array ties
* must be in the form {2,3,2,3} where each integer denotes the occurrence of a tie
* and the number of elements involved.
*
* if no ties, a void array shoud be used (i.e. ties=new int [0];)
*
* @param N n parameter
* @param R rank associated
* @param ties weight of ties
* @return p-value computed
*/
public double computeAsymptoticRightTailProbability(int N,double R,double ties){
double numerator2,denominator;
double z;
double rightTail;
normal=new NormalDistribution();
numerator2=R-0.5-(N*(N+1.0)/4.0);
denominator=Math.sqrt(N*(N+1.0)*((2*N)+1.0)/24.0);
denominator-=ties/48.0;
z=numerator2/denominator;
rightTail=1.0-normal.getTipifiedProbability(z, false);
return rightTail;
}//end-method
/**
* Computes asymptotic double tailed p-value of the Wilcoxon distribution
*
* A ties correction of variance is incorporated. The array ties
* must be in the form {2,3,2,3} where each integer denotes the occurrence of a tie
* and the number of elements involved.
*
* if no ties, a void array shoud be used (i.e. ties=new int [0];)
*
* @param N n parameter
* @param R rank associated
* @param ties weight of ties
* @return p-value computed
*/
public double computeAsymptoticLeftTailProbability(int N,double R,double ties){
double numerator1,denominator;
double z;
double leftTail;
normal=new NormalDistribution();
numerator1=R+0.5-(N*(N+1.0)/4.0);
denominator=Math.sqrt(N*(N+1.0)*((2*N)+1.0)/24.0);
denominator-=ties/48.0;
z=numerator1/denominator;
leftTail=normal.getTipifiedProbability(z, false);
return leftTail;
}//end-method
}//end-class
Ads help maintain this website.