public void GetSimilarityStringsDefault()
{
var bagOfTokensSim = new BagOfWordsSimilarity();
var patternTokens = new[] { "anna", "kurnikovova" };
var targetTokens = new[] { "kurnikovova", "anna" };
var scoreRes = bagOfTokensSim.GetSimilarity(patternTokens, targetTokens);
scoreRes.Should().Be(1.0);
var patternTokens2 = new[] { "jaromir", "jagr" };
var targetTokens2 = new[] { "jaromir", "jager" };
var scoreRes2 = bagOfTokensSim.GetSimilarity(patternTokens2, targetTokens2);
scoreRes2.Should().Be(0.9);
}
public void GetSimilarityStringsLevenshtein()
{
var bagOfTokensSim = new BagOfWordsSimilarity(TokenSimilarity.Levenshtein);
bagOfTokensSim.InternalTokenSimilarity.Should().Be(TokenSimilarity.Levenshtein);
var patternTokens = new[] { "anna", "kurnikovova" };
var targetTokens = new[] { "kurnikovova", "anna" };
var scoreRes = bagOfTokensSim.GetSimilarity(patternTokens, targetTokens);
scoreRes.Should().Be(1.0);
var patternTokens2 = new[] { "jaromir", "jagr" };
var targetTokens2 = new[] { "jaromir", "jager" };
var scoreRes2 = bagOfTokensSim.GetSimilarity(patternTokens2, targetTokens2);
scoreRes2.Should().BeApproximately(0.9, ErrorTolerance);
}
public void GetSimilarityNormalizedStringNotSymmetric()
{
var simNormString = new BagOfWordsSimilarity();
var patternTokens = new[] { new NormalizedString("miss"), new NormalizedString("anna"), new NormalizedString("kurnikovova") };
var targetTokens = new[] { new NormalizedString("kurnikovova"), new NormalizedString("anna"), };
var simResult = simNormString.GetSimilarity(patternTokens, targetTokens);
simResult.Should().BeApproximately(0.696, ErrorTolerance);
}
public void GetSimilarityTokenizerNotSymmetric()
{
const string pattern = "miss anna kurnikovova";
const string target = "kurnikovova anna";
var simNormString = new BagOfWordsSimilarity();
var patternTokenizer = new Tokenizer(pattern);
var targetTokenizer = new Tokenizer(target);
var simResult = simNormString.GetSimilarity(patternTokenizer, targetTokenizer);
simResult.Should().BeApproximately(0.696, ErrorTolerance);
}
public void GetSimilaritySymmetric()
{
// not symmetric instance of BagOfTokensSimilarity
var bagOfTokensNotSymmetric = new BagOfWordsSimilarity(TokenSimilarity.Levenshtein, false);
bagOfTokensNotSymmetric.IsSymmetric.Should().Be(false);
// symmetric instance of BagOfTokensSimilarity
var bagOfTokensSymmetric = new BagOfWordsSimilarity(TokenSimilarity.Levenshtein, true);
bagOfTokensSymmetric.IsSymmetric.Should().Be(true);
var patternTokens = new[] { "miss", "anna", "kurnikovova" };
var targetTokens = new[] { "kurnikovova", "anna" };
var scoreSymmetricRes = bagOfTokensNotSymmetric.GetSimilarity(patternTokens, targetTokens);
scoreSymmetricRes.Should().BeApproximately(0.696, ErrorTolerance);
var scoreNotSymmetricRes = bagOfTokensSymmetric.GetSimilarity(patternTokens, targetTokens);
scoreNotSymmetricRes.Should().Be(1.0);
}
private static void Main(string[] args)
{
/****************************************************/
/* BagOfTokensSimilarity */
/****************************************************/
// the recommend method for complex similarity on more words
var bagOfTokens = new BagOfWordsSimilarity();
const string pattern = "John Smith";
const string targetText = "Mr. John Smith, Jr.";
// using normalized string and tokenizer returns score 1.0
var resultingSim = bagOfTokens.GetSimilarity(new Tokenizer(new NormalizedString(pattern)),
new Tokenizer(new NormalizedString(targetText)));
/****************************************************/
/* Levenshtein */
/****************************************************/
const string nameCorrect = "martha";
const string nameError = "marhta";
// Levenshtein distance (implements interface IDistance)
// & similarity (implements interface ISimilarity)
var lev = new Levenshtein();
// returns edit distance 2
var distLev = lev.GetDistance(nameCorrect, nameError);
// normalized string removes special symbols, diacritics and make case insensitivity
// returns score 0.67
var simLev = lev.GetSimilarity(new NormalizedString(nameCorrect), new NormalizedString(nameError));
/****************************************************/
/* Damerau-Levenshtein */
/****************************************************/
// DamerauLevenshtein implements IDistance and ISimilarity
var damLev = new DamerauLevenshtein();
// returns edit distance 1
var distDamLev = damLev.GetDistance(nameCorrect, nameError);
// returns score 0.83
var simDamLev = damLev.GetSimilarity(nameCorrect, nameError);
/****************************************************/
/* Jaro, Jaro-Winler */
/****************************************************/
// Jaro, Jaro-Winkler implements only ISimilarity
var nameFirst = new Token("dwayne");
var nameSecond = new Token("duane");
var jaro = new Jaro();
var jaroWinkler = new JaroWinkler();
jaroWinkler.GetSimilarity(nameFirst, nameSecond);
/****************************************************/
/* Q-grams coefficient */
/****************************************************/
// q-grams similarity coefficient - Dice, Jaccard, Overlap
// with different q-grams type
var diceUnigrams = new DiceCoefficient <Bigram>();
var jaccardBigrams = new JaccardCoefficient <Bigram>();
var overlapTrigrams = new OverlapCoefficient <Bigram>();
// returns score 0.5
var jaccardSim = jaccardBigrams.GetSimilarity(pattern, targetText);
// returns score 0.67
var diceSim = diceUnigrams.GetSimilarity(pattern, targetText);
// returns score 1.0
var overlapSim = overlapTrigrams.GetSimilarity(pattern, targetText);
Console.ReadKey();
}