de.linguatools.disco

Class KolbVectorSimilarity

java.lang.Object

de.linguatools.disco.KolbVectorSimilarity

All Implemented Interfaces:

VectorSimilarity

public class KolbVectorSimilarity
extends java.lang.Object
implements VectorSimilarity

Constructor Summary

Constructors

Constructor and Description
KolbVectorSimilarity()

Method Summary

Modifier and Type	Method and Description
double	computeSimilarity(org.apache.lucene.document.Document doc1, org.apache.lucene.document.Document doc2) This method compares two word vectors from a DISCOLuceneIndex using the similarity measure SimilarityMeasures.KOLB that is described in the paper Peter Kolb.
double	computeSimilarity(float[] denseVector1, float[] denseVector2) Compute similarity between two dense vectors.
double	computeSimilarity(java.util.Map<java.lang.String,java.lang.Float> mapVector1, java.util.Map<java.lang.String,java.lang.Float> mapVector2) Compute similarity between two sparse vectors.

Methods inherited from class java.lang.Object

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

Constructor Detail

KolbVectorSimilarity

public KolbVectorSimilarity()

Method Detail

computeSimilarity

public double computeSimilarity(float[] denseVector1,
                                float[] denseVector2)

Description copied from interface: VectorSimilarity

Compute similarity between two dense vectors.

Specified by:

computeSimilarity in interface VectorSimilarity

Returns:

computeSimilarity

public double computeSimilarity(org.apache.lucene.document.Document doc1,
                                org.apache.lucene.document.Document doc2)

This method compares two word vectors from a DISCOLuceneIndex using the similarity measure SimilarityMeasures.KOLB that is described in the paper

Peter Kolb. Experiments on the difference between semantic similarity and relatedness. In Proceedings of the 17th Nordic Conference on Computational Linguistics - NODALIDA '09, Odense, Denmark, May 2009.

Specified by:

computeSimilarity in interface VectorSimilarity

Parameters:

doc1 - A document from the Lucene index containing the word vector for word #1

doc2 - A document from the Lucene index containing the word vector for word #2

Returns:

the similarity between the two word vectors; a value between 0.0F and 1.0F.

computeSimilarity

public double computeSimilarity(java.util.Map<java.lang.String,java.lang.Float> mapVector1,
                                java.util.Map<java.lang.String,java.lang.Float> mapVector2)

Description copied from interface: VectorSimilarity

Compute similarity between two sparse vectors.

Specified by:

computeSimilarity in interface VectorSimilarity

Returns: