public static double findSimilarDictionaryWord(string word, double maxSimilarity, int index, List<string> equalMinDistanceDictWordList)
{
index = index - _minWordLength;
word = word.ToLower();
double NewSimilarity = 0;
int WordLength = word.Length;
if ((WordLength + index) < 0)
return maxSimilarity;
if ((WordLength + index) >= _IndexDictionary.Length)
return maxSimilarity;
if (_IndexDictionary[WordLength + index] == null)
return maxSimilarity;
for (int j = 0; j < _IndexDictionary[WordLength + index].Count; j++)
{
JaroWinklerDistance JaroDist = new JaroWinklerDistance();
NGramDistance ng = new NGramDistance();
JaccardDistance jd = new JaccardDistance();
NewSimilarity = jd.GetDistance(word, _IndexDictionary[WordLength + index][j]);//(double)JaroDist.GetDistance(word, _IndexDictionary[WordLenght - 1 + index][j]);
if (NewSimilarity > maxSimilarity)
{
equalMinDistanceDictWordList.Clear();
equalMinDistanceDictWordList.Add(_IndexDictionary[WordLength + index][j]);
maxSimilarity = NewSimilarity;
}
else if (NewSimilarity == maxSimilarity)
equalMinDistanceDictWordList.Add(_IndexDictionary[WordLength + index][j]);
}
return maxSimilarity;
}
public void TestGetDistance2()
{
StringDistance sd = new NGramDistance(2);
float d = sd.GetDistance("al", "al");
Assert.AreEqual(d, 1.0f, 0.001);
d = sd.GetDistance("a", "a");
Assert.AreEqual(d, 1.0f, 0.001);
d = sd.GetDistance("b", "a");
Assert.AreEqual(d, 0.0f, 0.001);
d = sd.GetDistance("a", "aa");
Assert.AreEqual(d, 0.5f, 0.001);
d = sd.GetDistance("martha", "marhta");
Assert.AreEqual(d, 0.6666, 0.001);
d = sd.GetDistance("jones", "johnson");
Assert.AreEqual(d, 0.4285, 0.001);
d = sd.GetDistance("natural", "contrary");
Assert.AreEqual(d, 0.25, 0.001);
d = sd.GetDistance("abcvwxyz", "cabvwxyz");
Assert.AreEqual(d, 0.625, 0.001);
d = sd.GetDistance("dwayne", "duane");
Assert.AreEqual(d, 0.5833, 0.001);
d = sd.GetDistance("dixon", "dicksonx");
Assert.AreEqual(d, 0.5, 0.001);
d = sd.GetDistance("six", "ten");
Assert.AreEqual(d, 0, 0.001);
float d1 = sd.GetDistance("zac ephron", "zac efron");
float d2 = sd.GetDistance("zac ephron", "kai ephron");
Assert.IsTrue(d1 > d2);
d1 = sd.GetDistance("brittney spears", "britney spears");
d2 = sd.GetDistance("brittney spears", "brittney startzman");
Assert.IsTrue(d1 > d2);
d1 = sd.GetDistance("0012345678", "0012890678");
d2 = sd.GetDistance("0012345678", "0072385698");
Assert.AreEqual(d1, d2, 0.001);
}
public static double findSimilarDictionaryWord(string word, double maxSimilarity, int index, Dictionary<string, double> equalMinDistanceDictWordList)
{
try
{
double distancethreshold = 0.3;
index = index - _minWordLength;
double NewDistance = 0;
int WordLenght = word.Length;
if ((WordLenght + index) < 0)
return maxSimilarity;
if ((WordLenght + index) >= _IndexDictionary.Length)
return maxSimilarity;
if (_IndexDictionary[WordLenght - 1 + index] == null)
return 0;
for (int j = 0; j < _IndexDictionary[WordLenght - 1 + index].Count; j++)
{
JaroWinklerDistance JaroDist = new JaroWinklerDistance();
NGramDistance ng = new NGramDistance();
JaccardDistance jd = new JaccardDistance();
string temp = _IndexDictionary[WordLenght - 1 + index][j];
NewDistance = jd.GetDistance(word, temp);
double NewDistance2 = -1;
if (NewDistance < NewDistance2)
NewDistance = NewDistance2;
if (NewDistance > maxSimilarity)
{
foreach (var item in equalMinDistanceDictWordList.ToList())
{
if (item.Value <= NewDistance - distancethreshold)
equalMinDistanceDictWordList.Remove(item.Key);
}
tempReplacement = temp;
if (!equalMinDistanceDictWordList.ContainsKey(temp))
equalMinDistanceDictWordList.Add(temp, NewDistance);
//else
// equalMinDistanceDictWordList[tempReplacement] = NewDistance;
maxSimilarity = NewDistance;
}
else if (NewDistance <= maxSimilarity + distancethreshold && NewDistance >= maxSimilarity - distancethreshold && NewDistance > 0)
if (!equalMinDistanceDictWordList.ContainsKey(temp))
equalMinDistanceDictWordList.Add(temp, NewDistance);
}
}
catch (Exception e)
{
throw e;
}
return maxSimilarity;
}
public static double findSimilarDictionaryWord(string word, double maxSimilarity, int index, List<string> equalMinDistanceDictWordList, bool exact)
{
index = index - _minWordLength;
int WordLength = word.Length;
int index2 = index;
if (index < 0 || (WordLength>=2 && char.IsUpper(word[0])&&!char.IsUpper(word[1])) )
index2 = 0;
word = word.ToLower();
bool noSpace = false;
if (word.CompareTo(word.Trim()) == 0)
noSpace = true;
else
word = word.Trim();
double NewSimilarity = 0;
if ((WordLength + index) < 0)
return maxSimilarity;
if ((WordLength + index) >= _IndexDictionary.Length)
return maxSimilarity;
if (_IndexDictionary[WordLength + index] == null)
return maxSimilarity;
for (int j = 0; j < _IndexDictionary[WordLength + index].Count; j++)
{
JaroWinklerDistance JaroDist = new JaroWinklerDistance();
NGramDistance ng = new NGramDistance();
JaccardDistance jd = new JaccardDistance();
string temp = _IndexDictionary[WordLength + index][j];
if(noSpace&&temp.CompareTo(word)==0)
{
equalMinDistanceDictWordList.Clear();
equalMinDistanceDictWordList.Add(temp);
return 10;
}
else if (temp.Contains(word))
{
equalMinDistanceDictWordList.Add(/*item);*/temp);
maxSimilarity = 1;
}
else if(index <= 2)
{
for (int i = 0; i <= index2; i++)
{
string s = temp.Substring(i);
string s2 = temp.Substring(0, temp.Length - index2);
//Console.WriteLine(item);
if (!exact)
NewSimilarity = Math.Max(jd.GetDistance(word, s), jd.GetDistance(word, s2));
else
{
NewSimilarity = jd.GetDistance(word, temp);
if (NewSimilarity == 1)
{
equalMinDistanceDictWordList.Clear();
equalMinDistanceDictWordList.Add(s);
maxSimilarity = NewSimilarity;
}
return maxSimilarity;
}
if (NewSimilarity > .33)
{
//equalMinDistanceDictWordList.Clear();
equalMinDistanceDictWordList.Add(/*item);*/temp);
maxSimilarity = NewSimilarity;
break;
}
}
}
}
return maxSimilarity;
}
public static void TestQGram()
{
string fn1 = "acht";
string fn2 = "yacht";
string fn3 = "acha";
NGramDistance ngd = new NGramDistance();
double x = ngd.GetDistance(fn1, fn2);
double y = ngd.GetDistance(fn1, fn3);
double z = x;
}
public void TestEmpty()
{
StringDistance nsd = new NGramDistance(1);
float d = nsd.GetDistance("", "al");
Assert.AreEqual(d, 0.0f, 0.001);
}
