ws4j 中句子级别的语义匹配

Posted 2023-02-24

【中文标题】ws4j 中句子级别的语义匹配

【英文标题】：Semantic matching in ws4j at sentence level

【发布时间】：2016-07-02 04:08:14

【问题描述】：

我目前正在尝试在语义上匹配 ws4j 中的两个句子。我在单词级别实现了这个概念，但在句子级别实现同样的概念时遇到了麻烦，并以矩阵形式获得输出，就像它在在线演示中显示的那样。如何开发代码来做同样的事情？

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);
    

【问题讨论】：

请提供详细信息，展示您为帮助他人做出贡献所做的努力。

***.com/users/3639557/user3639557 我附上了我实现的 API。它为单词生成正确的值，但为句子生成垃圾值

【参考方案1】：

我遇到了类似的问题，这个例子有效：

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);