it.uniroma2.sag.kelp.data.representation.structure.similarity

Class LexicalStructureElementSimilarity

java.lang.Object

it.uniroma2.sag.kelp.data.representation.structure.similarity.VectorBasedStructureElementSimilarity

it.uniroma2.sag.kelp.data.representation.structure.similarity.LexicalStructureElementSimilarity

All Implemented Interfaces:

StructureElementSimilarityI

Direct Known Subclasses:

CompositionalNodeSimilarity

public class LexicalStructureElementSimilarity
extends VectorBasedStructureElementSimilarity

This class implements a similarity function between StructureElements. A strict label matching is applied on all structure elements whose type differs LexicalStructureElement. Similarity between LexicalStructureElements is estimated by projecting represented words in a WordSpace and evaluating a Kernel function between these vectors. By default a dot product provided by a LinearKernel is estimated.

Author:

Danilo Croce, Simone Filice

Field Summary

Fields inherited from class it.uniroma2.sag.kelp.data.representation.structure.similarity.VectorBasedStructureElementSimilarity

VECTOR_NAME

Constructor Summary

Constructors

Constructor and Description
LexicalStructureElementSimilarity()
LexicalStructureElementSimilarity(WordspaceI wordspace)

Method Summary

Methods

Modifier and Type	Method and Description
boolean	getAllowDifferentPOS() Returns whether the similarity between words having different Part-of-Speech is allowed or if it must be set to 0
boolean	getIgnorePosInLemmaMatches() Returns whether two lexical structure elements must provide a perfect match if their lemmas are the same, regardless their part-of-speeches
void	setAllowDifferentPOS(boolean allowDifferentPOS) Sets whether the similarity between words having different Part-of-Speech is allowed or if it must be set to 0
void	setIgnorePosInLemmaMatches(boolean ignorePosInLemmaMatches) Sets whether two lexical structure elements must provide a perfect match if their lemmas are the same, regardless their part-of-speeches
float	sim(StructureElement sx, StructureElement sz) This function measure the similarity between structure elements

Methods inherited from class it.uniroma2.sag.kelp.data.representation.structure.similarity.VectorBasedStructureElementSimilarity

getCorrespondingVector, getEnrichmentName, getKernel, getSimilarity, getWordspace, setEnrichmentName, setKernel, setWordspace

Methods inherited from class java.lang.Object

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

Constructor Detail

LexicalStructureElementSimilarity

public LexicalStructureElementSimilarity()

LexicalStructureElementSimilarity

public LexicalStructureElementSimilarity(WordspaceI wordspace)
                                  throws NumberFormatException,
                                         IOException

Parameters:

matrixPath - The path of the wordSpace to be loaded. By default the wordpspace will be of the class Wordpspace

Throws:

NumberFormatException

IOException

Method Detail

sim

public float sim(StructureElement sx,
        StructureElement sz)

Description copied from interface: StructureElementSimilarityI

This function measure the similarity between structure elements

Parameters:

sx - the first structure element

sz - the second structure element

Returns:

the similarity between two structure elements

getAllowDifferentPOS

public boolean getAllowDifferentPOS()

Returns whether the similarity between words having different Part-of-Speech is allowed or if it must be set to 0

Returns:

whether the similarity between words having different Part-of-Speech is allowed

setAllowDifferentPOS

public void setAllowDifferentPOS(boolean allowDifferentPOS)

Sets whether the similarity between words having different Part-of-Speech is allowed or if it must be set to 0

Parameters:

allowDifferentPOS -

getIgnorePosInLemmaMatches

public boolean getIgnorePosInLemmaMatches()

Returns whether two lexical structure elements must provide a perfect match if their lemmas are the same, regardless their part-of-speeches

Returns:

whether two lexical structure elements must provide a perfect match if their lemmas are the same, regardless their part-of-speeches

setIgnorePosInLemmaMatches

public void setIgnorePosInLemmaMatches(boolean ignorePosInLemmaMatches)

Sets whether two lexical structure elements must provide a perfect match if their lemmas are the same, regardless their part-of-speeches

Parameters:

ignorePosInLemmaMatches - whether two lexical structure elements must provide a perfect match if their lemmas are the same, regardless their part-of-speeches