sentencesimilarity.cs

是用c#编写的

/*
 Measure semantic similarity between two sentences
 (disregards PartOfSpeech tagging and WordSenseDisambiguation)
 Author: Thanh Ngoc Dao - [e-mail]
 Copyright (c) 2005 by Thanh Ngoc Dao.
*/

using System;

namespace WordsMatching
{
	/// <summary>
	/// Measuring relationship between two given sentences
	/// </summary>
	public class SentenceSimilarity
	{
        //private int[] _senses1, _senses2;
        //float[,] _similarity;

		string[] _source, _target;
		private int m, n;

		public SentenceSimilarity()
		{
			StopWordsHandler stopword=new StopWordsHandler() ;
		}

        private MyWordInfo[] Disambiguate(string[] words)
        {
            if (words.Length == 0) return null;

            MyWordInfo[] wordInfos=new MyWordInfo [words.Length];
            
            for (int i = 0; i < words.Length; i++)
            {
                
                WnLexicon.WordInfo wordInfo = WnLexicon.Lexicon.FindWordInfo(words[i], true);

                if (wordInfo.partOfSpeech != Wnlib.PartsOfSpeech.Unknown)
                {
                    if (wordInfo.text != string.Empty)
                        words[i] = wordInfo.text;

                    Wnlib.PartsOfSpeech[] posEnum = (Wnlib.PartsOfSpeech[])Enum.GetValues(typeof(Wnlib.PartsOfSpeech));

                    for (int j = 0; j < posEnum.Length; j++)
                    {
                        if (wordInfo.senseCounts[j] > 0) // get the first part of speech
                        {
                            wordInfos[i] = new MyWordInfo(words[i], posEnum[j]);                             
                            break;
                        }
                    }
                }
            }

            WordSenseDisambiguator wsd = new WordSenseDisambiguator();
            wordInfos=wsd.Disambiguate(wordInfos);

            return wordInfos;
        }


        //MyWordInfo[] _myWordsInfo_i, _myWordsInfo_j;        
        //private void MyInitOld()
        //{
        //    _myWordsInfo1 = Disambiguate(_source);
        //    _myWordsInfo2 = Disambiguate(_target);

        //    m = _myWordsInfo1.Length; n = _myWordsInfo2.Length;
        //    _similarity =new float[m, n] ;

        //    for (int i=0; i < m; i++)
        //    {
        //        _myWordsInfo1[i].Sense = _myWordsInfo1[i].Sense < 0 ? 0 : _myWordsInfo1[i].Sense;                

        //        string word1 = _source[i];
        //        for (int j=0; j < n; j++)
        //        {
        //            _myWordsInfo2[i].Sense = _myWordsInfo2[i].Sense < 0 ? 0 : _myWordsInfo2[i].Sense;					

        //            string word2=_target[j];
        //            WordDistance distance = new WordDistance();
        //            float weight = distance.GetSimilarity(_myWordsInfo1[i], _myWordsInfo2[j]);					

        //            _similarity[i, j]=weight;					
        //        }
        //    }
        //}



        float[][] GetSimilarityMatrix(string[] string1, string[] string2)
        {
            m = string1.Length; n = string2.Length;            
            float[][] simMatrix = new float[m][];            
            
            Wnlib.PartsOfSpeech[] POSEnum = (Wnlib.PartsOfSpeech[])Enum.GetValues(typeof(Wnlib.PartsOfSpeech));
        	HierarchicalWordData[][] wordData_1 = new HierarchicalWordData[m][];
        	HierarchicalWordData[][] wordData_2 = new HierarchicalWordData[n][];
            for (int i = 0; i < m; i++) 
                simMatrix[i] = new float[n];

            for (int i = 0; i < m; i++)
                wordData_1[i] = new HierarchicalWordData[POSEnum.Length];
            for (int j = 0; j < n; j++)
                wordData_2[j] = new HierarchicalWordData[POSEnum.Length];

            for (int i = 0; i < m; i++)             
            {                                                                                
                for (int j = 0; j < n; j++)
                {
                    float synDist = AcronymChecker.GetEditDistanceSimilarity(string1[i], string2[j]);

                    for (int partOfSpeech = 1; partOfSpeech < POSEnum.Length; partOfSpeech++)
                    {
                         if (wordData_1[i][partOfSpeech] == null)
                         {
                             MyWordInfo myWordsInfo_i = new MyWordInfo(string1[i], POSEnum[partOfSpeech]);
                             wordData_1[i][partOfSpeech] = new HierarchicalWordData(myWordsInfo_i);
                         }
                         if (wordData_2[j][partOfSpeech] == null)
                         {
                             MyWordInfo myWordsInfo_j = new MyWordInfo(string2[j], POSEnum[partOfSpeech]);
                             wordData_2[j][partOfSpeech] = new HierarchicalWordData(myWordsInfo_j);
                         }

                         WordSimilarity wordDistance = new WordSimilarity();
                         float semDist = wordDistance.GetSimilarity(wordData_1[i][partOfSpeech], wordData_2[j][partOfSpeech]);
                         float weight = Math.Max(synDist, semDist);
                         if (simMatrix[i][j] < weight)
                             simMatrix[i][j] = weight;                    
                    }
                }                                    
            }            
         
         return simMatrix;
      }

        public float GetScore(string string1, string string2)		
		{			
			Tokeniser tok=new Tokeniser() ;
            tok.UseStemming = false;

			_source=tok.Partition(string1) ;
			_target=tok.Partition(string2) ;

			if (_source.Length == 0 || _target.Length == 0 )
				return 0F;

            float[][] simMatrix = GetSimilarityMatrix(_source, _target);		
			HeuristicMatcher match=new HeuristicMatcher() ;
            //float score = HeuristicMatcher.ComputeSetSimilarity(simMatrix, 2, 0.3F);
            float score = HeuristicMatcher.ComputeSetSimilarity(simMatrix, 1);
			return score;	
		}
	}
}