com.articulate.sigma.nlp

Class WNsim

java.lang.Object

com.articulate.sigma.nlp.WNsim

public class WNsim
extends Object

an implementation of Resnik, P. (1995). Using information content to evaluate semantic similarity in a taxonomy. In Proceedings of the 14th International Joint Conference on Artificial Intelligence, 448–453. Adding a similarity method based on synset subsumption.

Constructor Summary

Constructors

Constructor and Description
WNsim()

Method Summary

Modifier and Type	Method and Description
static void	calcMaxFreqs()
static HashMap<String,Integer>	computeSubsumingFreq(HashSet<String> childRels, HashSet<String> parentRels)
static int	computeSubsumingFreq2(HashSet<String> childRels, HashSet<String> parentRels)
static int	computeSubsumingFreq2(String synset, HashSet<String> childRels, HashSet<String> parentRels)
static float	computeSynsetSim(String s1, String s2) Compute similarity of two synsets.
static float	computeWordSim(String s1, String s2, boolean allSynsets) Compute similarity of two words as the maximum similarity among their possible synsets.
static double	correlation(HashMap<String,Float> x, HashMap<String,Float> y) determine the sample correlation coefficient
static void	createIndexes()
static float	deviation(HashMap<String,Float> x, HashMap<String,Float> y, float meanx, float meany)
static HashSet<String>	getChildList(String s)
static int	getIndex(String s)
static HashMap<String,Integer>	getSubsumingFreq()
static void	main(String[] args) A main method, used only for testing.
static float	maxCommonSynsetOverlap(String s1, String s2)
static float	mean(HashMap<String,Float> x)
static HashMap<String,HashMap<String,Float>>	readTestFile(String filename, String regex, boolean removePOS) Read a space-delimited file of word pairs and their similarity rating.
static BitSet	setBits(String s)
static float	sqDeviation(HashMap<String,Float> x, float mean)
static int	synsetOverlap(String s1, String s2) Compute the number of child synsets shared by the two synsets
static float	synsetPercentOverlap(String s1, String s2) Compute the percentage of child synsets of the first synset shared by the second synset
static void	test()

Methods inherited from class java.lang.Object

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

Constructor Detail

WNsim

public WNsim()

Method Detail

calcMaxFreqs

public static void calcMaxFreqs()

getSubsumingFreq

public static HashMap<String,Integer> getSubsumingFreq()

computeSubsumingFreq

public static HashMap<String,Integer> computeSubsumingFreq(HashSet<String> childRels,
                                                           HashSet<String> parentRels)

computeSubsumingFreq2

public static int computeSubsumingFreq2(HashSet<String> childRels,
                                        HashSet<String> parentRels)

computeSubsumingFreq2

public static int computeSubsumingFreq2(String synset,
                                        HashSet<String> childRels,
                                        HashSet<String> parentRels)

computeSynsetSim

public static float computeSynsetSim(String s1,
                                     String s2)

Compute similarity of two synsets. Note that currently only synsets of the same part of speech have any similarity score.

computeWordSim

public static float computeWordSim(String s1,
                                   String s2,
                                   boolean allSynsets)

Compute similarity of two words as the maximum similarity among their possible synsets. Note that currently only synsets of the same part of speech have any similarity score.

Parameters:

allSynsets - determines whether to take the best fit from all possible synsets for the words, or just to compare the most frequent synsets for each word

readTestFile

public static HashMap<String,HashMap<String,Float>> readTestFile(String filename,
                                                                 String regex,
                                                                 boolean removePOS)

Read a space-delimited file of word pairs and their similarity rating.

mean

public static float mean(HashMap<String,Float> x)

deviation

public static float deviation(HashMap<String,Float> x,
                              HashMap<String,Float> y,
                              float meanx,
                              float meany)

sqDeviation

public static float sqDeviation(HashMap<String,Float> x,
                                float mean)

correlation

public static double correlation(HashMap<String,Float> x,
                                 HashMap<String,Float> y)

determine the sample correlation coefficient

getChildList

public static HashSet<String> getChildList(String s)

Returns:

a set of all the immediate children of the given synset from hyponym and instance hyponym links

getIndex

public static int getIndex(String s)

Parameters:

s - is a POS-prefixed synset

Returns:

an integer index into the bit vector

setBits

public static BitSet setBits(String s)

Parameters:

s - is a POS-prefixed synset

Returns:

a bit vector comprising set bits for all the children of the given synset

createIndexes

public static void createIndexes()

synsetOverlap

public static int synsetOverlap(String s1,
                                String s2)

Compute the number of child synsets shared by the two synsets

synsetPercentOverlap

public static float synsetPercentOverlap(String s1,
                                         String s2)

Compute the percentage of child synsets of the first synset shared by the second synset

maxCommonSynsetOverlap

public static float maxCommonSynsetOverlap(String s1,
                                           String s2)

test

public static void test()

main

public static void main(String[] args)

A main method, used only for testing. It should not be called during normal operation.