private static double CalculateSimilarity(List<int> wordId1, List<int> wordId2)
{
if (lw1 != null && lw2 != null)
{
int universeSize = (lw1.Count + lw2.Count) / 2;
int numHashFunctions = 100;
using (minHash = new MinHash(universeSize, numHashFunctions))
{
List<uint> mh1 = minHash.GetMinHash(wordId1);
List<uint> mh2 = minHash.GetMinHash(wordId2);
return minHash.Similarity(mh1, mh2);
}
}
else throw new NullReferenceException("Fill the word lists first before continuing to proceed similarity calculation.");
}
protected void saveButton_Click(object sender, EventArgs e)
{
//hashing no. + hasil akhir
List<double> jaccard = new List<double>();
List<double> pearson = new List<double>();
//tokenize
var lw1 = TFIDF.Tokenize(teks1.Text);
var lw2 = TFIDF.Tokenize(teks2.Text);
int hashno = Convert.ToInt32(Math.Ceiling((decimal)(lw1.Count+ lw2.Count)/2));
MinHash mh = new MinHash(Math.Min(lw1.Count, lw2.Count), hashno);
//vocab list
var vl1 = new List<string>();
var vl2 = new List<string>();
//pasangan antara ID kamus dengan string yang dicocokkan
Dictionary<int, string> res1 = new Dictionary<int, string>();
Dictionary<int, string> res2 = new Dictionary<int, string>();
//list ID kamus aja
List<int> sr1 = wn.GetSemanticRelation(lw1, 1, 2, out vl1, out res1, dictfile, stemmer);
List<int> sr2 = wn.GetSemanticRelation(lw2, 2, 2, out vl2, out res2, dictfile, stemmer);
List<uint> hash1 = mh.GetMinHash(sr1);
List<uint> hash2 = mh.GetMinHash(sr2);
List<double> hashdb1 = new List<double>();
List<double> hashdb2 = new List<double>();
//convert uint to double
string hr1 = string.Empty;
string hr2 = string.Empty;
//print code + id
foreach (var item in res1)
{
hr1 += "<p>" + item.Value + ": " + item.Key.ToString() + "</p>";
}
foreach (var item in res2)
{
hr2 += "<p>" + item.Value + ": " + item.Key.ToString() + "</p>";
}
foreach (var item in hash1)
{
var cvt = Convert.ToDouble(item);
hashdb1.Add(cvt);
}
foreach (var item in hash2)
{
var cvt = Convert.ToDouble(item);
hashdb2.Add(Convert.ToDouble(cvt));
}
//final jaccard & pearson
var jacc = ((double)mh.Similarity(hash1, hash2));
var jaccpercent = (double)jacc * 100;
var jaccmin = (double)(jacc - 1);
var pear = pr.Calc(hashdb1, hashdb2);
jaccardResult.Text = jaccpercent + "%";
pearsonResult.Text = (double)(100-pear) + "%";
hashresult1.Text = hr1;
hashresult2.Text = hr2;
divresult.Visible = true;
}
private static void ExportExcel(int currentindex)
{
Pearson pr = new Pearson();
int hashno = 100;
using (ExcelPackage ep = new ExcelPackage())
{
ep.Workbook.Properties.Title = "FINGERPRINT ANALYSIS";
ep.Workbook.Properties.Author = "THEODORUS YOGA M";
var ws = ep.Workbook.Worksheets.Add("Fingerprint Result");
ws.Cells[1, 1].Value = "Hash No.";
ws.Cells[1, 2].Value = "Jaccard";
ws.Cells[1, 3].Value = "Pearson";
ws.Cells[1, 4].Value = "Euclidean";
int rowstart = 2;
List<double> jaccard = new List<double>();
List<double> jaccardMin = new List<double>();
List<double> pearson = new List<double>();
List<double> euclidean = new List<double>();
while (hashno <= 5000)
{
//Console.WriteLine("Processing " + hashno + " hash functions");
MinHash mh = new MinHash(Math.Min(lw1.Count, lw2.Count), hashno);
//List<int> example1 = new List<int>() { 21, 50, 67, 101, 108 };
//List<int> example2 = new List<int>() { 23, 52, 90, 102, 110 };
hash1 = mh.GetMinHash(sr1); //get minimal hashing for 1st semantic relation
hash2 = mh.GetMinHash(sr2); //get minimal hashing for 2nd semantic relation
//hash1 = mh.GetMinHash(example1);
//hash2 = mh.GetMinHash(example2);
//convert to double for pearson
List<double> hashdb1 = new List<double>();
List<double> hashdb2 = new List<double>();
int count1 = hash1.Count;
int current = 1;
foreach (uint hash in hash1)
{
hashdb1.Add(Convert.ToDouble(hash));
//Console.SetCursorPosition(0, Wordnet.consoleln);
//Console.WriteLine("\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t");
//Console.SetCursorPosition(0, Wordnet.consoleln);
//Console.WriteLine("STATUS: Processing hash functions in document 1 of 2: " + current + "/" + count1);
current++;
}
Wordnet.consoleln++;
int count2 = hash2.Count;
current = 1;
foreach (uint hash in hash2)
{
hashdb2.Add(Convert.ToDouble(hash));
//Console.SetCursorPosition(0, Wordnet.consoleln);
//Console.WriteLine("\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t");
//Console.SetCursorPosition(0, Wordnet.consoleln);
//Console.WriteLine("STATUS: Processing hash functions in document 2 of 2: " + current + "/" + count2);
current++;
}
Wordnet.consoleln++;
//calculate TF-IDF
List<List<string>> vocabsInDoc = new List<List<string>>();
vocabsInDoc.Add(vl1);
vocabsInDoc.Add(vl2);
var tfidf = TFIDF.CalculateTFIDF(vocabsInDoc);
Wordnet.consoleln++;
//Console.WriteLine("JACCARD SIMILARITY DEGREE: " + mh.Similarity(hash1, hash2));
//Console.WriteLine("PEARSON SIMILARITY DEGREE: " + pr.Calc(hashdb1, hashdb2));
//Console.WriteLine("TF-IDF EUCLIDEAN DISTANCE:" + TFIDF.EuclideanDist(tfidf[0], tfidf[1]));
Console.WriteLine(hashno + "\t,\t" + ((double)mh.Similarity(hash1, hash2)) + "\t,\t" +
pr.Calc(hashdb1, hashdb2) + "\t,\t" + TFIDF.EuclideanDist(tfidf[0], tfidf[1]));
ws.Cells[rowstart, 1].Value = hashno.ToString();
var jacc = ((double)mh.Similarity(hash1, hash2));
var jaccmin = (double)(jacc - 1);
var pear = pr.Calc(hashdb1, hashdb2);
var euc = TFIDF.EuclideanDist(tfidf[0], tfidf[1]);
ws.Cells[rowstart, 2].Value = jacc;
jaccard.Add(jacc);
pearson.Add(pear);
euclidean.Add(euc);
jaccardMin.Add(Math.Abs(jaccmin));
ws.Cells[rowstart, 3].Value = pear;
ws.Cells[rowstart, 4].Value = TFIDF.EuclideanDist(tfidf[0], tfidf[1]);
//foreach (var vector in tfidf)
//{
// foreach (var val in vector)
// {
// Console.Write(val + ";");
// }
// Console.WriteLine();
//}
hashno += 50;
rowstart++;
}
FileStream file;
if (!Directory.Exists("Out"))
Directory.CreateDirectory("Out");
if (!File.Exists("Out/naziefresult_" + currentindex.ToString() + ".xlsx"))
{
FileStream fs = File.Create("Out/naziefresult_" + currentindex.ToString() + ".xlsx");
fs.Dispose();
}
file = File.Open("Out/naziefresult_" + currentindex.ToString() + ".xlsx", FileMode.Open);
string append = string.Empty;
append += "JaccardMin<-c(";
foreach (var item in jaccardMin)
{
append += item.ToString().Replace(',', '.');
append += ", ";
}
append += ")";
append += "\n";
append += "Jaccard<-c(";
foreach (var item in jaccard)
{
append += item.ToString().Replace(',', '.');
append += ", ";
}
append += ")";
append += "\n";
append += "Pearson<-c(";
foreach (var item in pearson)
{
append += item.ToString().Replace(',', '.');
append += ", ";
}
append += ")";
append += "\n";
append += "Euclidean<-c(";
foreach (var item in euclidean)
{
append += item.ToString().Replace(',', '.');
append += ", ";
}
append += ")";
//Console.WriteLine("c(");
//foreach (var item in res)
//{
// Console.Write(item.ToString().Replace(',', '.'));
// Console.Write(", ");
//}
//Console.Write(")");
ep.SaveAs(file);
WriteFile(path + @"\Out\naziefr_" + currentindex.ToString() + ".txt", append);
file.Close();
}
}
