Semantic matching in ws4j at sentence level

1k views Asked by At

I am currently trying to semantically match two sentences in ws4j. I implemented the concept at a word level but am having trouble implementing the same at a sentence level and get an output in the form of a matrix like it shows on the online demo. How to develop a code to do the same?

import java.util.List;
import edu.cmu.lti.ws4j.impl.Lesk;
import edu.cmu.lti.jawjaw.pobj.POS;
import edu.cmu.lti.lexical_db.ILexicalDatabase;
import edu.cmu.lti.lexical_db.NictWordNet;
import edu.cmu.lti.lexical_db.data.Concept;
import edu.cmu.lti.ws4j.Relatedness;
import edu.cmu.lti.ws4j.RelatednessCalculator;

public class WordMatcher1 {
public static void main(String[] args)
{
    String word1="rifle";
    String word2="gun";

    ILexicalDatabase db = new NictWordNet();
    RelatednessCalculator lesk = new Lesk(db);

    List<POS[]> posPairs = lesk.getPOSPairs();
    double maxScore = -1D;

    for(POS[] posPair: posPairs) 
    {
        String p1 = null,p2 = null;
        List<Concept> synsets1 = (List<Concept>)db.getAllConcepts(word1, posPair[0].toString());
        List<Concept> synsets2 = (List<Concept>)db.getAllConcepts(word2, posPair[1].toString());

        for(Concept ss1: synsets1) 
        {
            for (Concept ss2: synsets2) 
            {
                p1 = ss1.getPos().toString();
                p2 = ss2.getPos().toString();
                Relatedness relatedness = lesk.calcRelatednessOfSynset(ss1, ss2);
                double score = relatedness.getScore();
                if (score > maxScore) 
                { 
                    maxScore = score;
                }

            }
        }

        if (maxScore == -1D) 
        {
            maxScore = 0.0;
        }
        System.out.println("sim('" + word1 +" "+ p1 +"', '" + p2 +" "+ word2+ "') =  " + maxScore);
    }
}
1

There are 1 answers

0
PaulDaigle On

I had a similar problem, and this example worked:

import java.util.List;
import edu.cmu.lti.jawjaw.pobj.POS;
import edu.cmu.lti.lexical_db.ILexicalDatabase;
import edu.cmu.lti.lexical_db.NictWordNet;
import edu.cmu.lti.lexical_db.data.Concept;
import edu.cmu.lti.ws4j.Relatedness;
import edu.cmu.lti.ws4j.RelatednessCalculator;
import edu.cmu.lti.ws4j.impl.Lesk;
import edu.cmu.lti.ws4j.util.WS4JConfiguration;

public class LeskSimilarity{

    public static void main(String[] args) {
    ILexicalDatabase db = new NictWordNet();
    RelatednessCalculator lesk = new Lesk(db);
    String word1="rifle";
    POS posWord1=  POS.n;
    String word2= "gun";
    POS posWord2= POS.n;
    double maxScore = 0;

        WS4JConfiguration.getInstance().setMFS(true);

        List<Concept> synsets1 = (List<Concept>)db.getAllConcepts(word1, posWord1.name());
        List<Concept> synsets2 = (List<Concept>)db.getAllConcepts(word2, posWord2.name());

        for(Concept synset1: synsets1) {
            for (Concept synset2: synsets2) {
                Relatedness relatedness =     lesk.calcRelatednessOfSynset(synset1, synset2);
            double score = relatedness.getScore();
            if (score > maxScore) { 
                maxScore = score;
            }
          }
        }

    if (maxScore == -1D) {
        maxScore = 0.0;
    }

    System.out.println("Similarity score of " + word1 + " & " + word2 + " : " + maxScore);
  }
}