Examples of GlossOverlapComparator

dkpro.similarity.algorithms.lsr.gloss.GlossOverlapComparator
Implements the distance measure by Lesk (1986). The distance between two terms a and b is defined as the number of shared tokens in the glosses of a and b. We normalize this value to fall into (0,1). @author zesch

Examples of dkpro.similarity.algorithms.lsr.gloss.GlossOverlapComparator

      return false;
    }


        try {
            boolean usePseudoGlosses = Boolean.parseBoolean(usePseudoGlossesString);
            measure = new GlossOverlapComparator(lsr, usePseudoGlosses);
        }
        catch (LexicalSemanticResourceException e) {
            throw new ResourceInitializationException(e);
        }

View Full Code Here

Examples of dkpro.similarity.algorithms.lsr.gloss.GlossOverlapComparator


  @Test
  public void testGlossOverlapWikipedia()
    throws Exception
  {
        LexSemResourceComparator comparator = new GlossOverlapComparator(wikiResource, false);


        Set<Entity> entitiesAuto  = wikiResource.getEntity("Automobile");
        Set<Entity> entitiesTruck = wikiResource.getEntity("Truck");


        assertTrue("Auto",  entitiesAuto.size() > 0);
        assertTrue("Truck", entitiesTruck.size() > 0);


        // same page
        assertEquals(1.0, comparator.getSimilarity(entitiesAuto, entitiesAuto), epsilon);


        // different pages
        assertEquals(0.2931, comparator.getSimilarity(entitiesAuto, entitiesTruck), epsilon);
    }

View Full Code Here

Examples of dkpro.similarity.algorithms.lsr.gloss.GlossOverlapComparator

    @Ignore("Currently deactivated as Wiktionary resource data is not for API version is not available.")
  @Test
  public void testGlossOverlapWiktionary()
    throws LexicalSemanticResourceException, SimilarityException
  {
    LexSemResourceComparator comparator = new GlossOverlapComparator(wktResource, false);


        Set<Entity> entities1  = wktResource.getEntity("Bayern");
        Set<Entity> entities2 = wktResource.getEntity("Deutschland");


        assertTrue("Bayern",  entities1.size() > 0);
        assertTrue("Deutschland", entities2.size() > 0);


        // same page
        assertEquals(1.0, comparator.getSimilarity(entities1, entities1), epsilon);


        // different pages
        assertEquals(0.22222, comparator.getSimilarity(entities1, entities2), epsilon);
    }

View Full Code Here

Examples of dkpro.similarity.algorithms.lsr.gloss.GlossOverlapComparator

  @Ignore("TODO for some reason the leskComparator gloss is from the disambiguation page - need to discover why")
  public void comparisonWithJwplImplementationTest()
    throws LexicalSemanticResourceException, WikiException, SimilarityException
  {
    wikiResource.setIsCaseSensitive(true);
    LexSemResourceComparator glossOverlapComparator = new GlossOverlapComparator(wikiResource, false);


        String term1 = "Automobile";
        String term2 = "Truck";


        Set<Entity> entitiesAuto  = wikiResource.getEntity(term1);
        Set<Entity> entitiesTruck = wikiResource.getEntity(term2);


        LeskComparator leskComparator = new LeskFirstComparator(wiki);
        leskComparator.setUseCache(false);


        assertEquals(glossOverlapComparator.getSimilarity(entitiesAuto, entitiesTruck), leskComparator.getSimilarity(term1, term2), epsilon);


        wikiResource.setIsCaseSensitive(false);
    }

View Full Code Here

TOP

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.