private static void Stemming_Setup()
{
stemmer = new Stemming();
moodList = new MoodListing().BingLiuMoodList;
bingliuTrie = new Trie <MoodValue>();
var bingLiuMoodList = new List <MoodItem>();
foreach (var item in moodList)
{
//var stemmed = stemmer.Stem(item.Word);
bingLiuMoodList.Add(new MoodItem {
Word = " " + item.Word + " ", Value = item.Value
});
}
bingLiuMoodList = bingLiuMoodList.Distinct().ToList();
foreach (var item in bingLiuMoodList)
{
bingliuTrie.Add(item.Word, item.Value);
}
bingliuTrie.Build();
drummondMoodList = new MoodListing().DrummondMoodList;
drummondTrie = new Trie <MoodValue>();
foreach (var item in drummondMoodList)
{
var stemmed = stemmer.Stem(item.Word);
drummondTrie.Add(" " + stemmed, item.Value);
}
drummondTrie.Build();
}
private void Treinar()
{
string cst = "Data Source=.\\SQLEXPRESS;AttachDbFilename=\"" + System.IO.Directory.GetCurrentDirectory() + "\\memorias\\test_memory.mdf\";Integrated Security=True;Connect Timeout=30;User Instance=True";
string document = txt_texto.Text.Trim();
Formatacao fmtc = new Formatacao();
Stemming stmg = new Stemming();
DataMining dm = new DataMining();
List <string> tokensFormatados = fmtc.Formatacao_completa(document);
List <string> tokensStemmizados = stmg.StemmizarListaDeTokens(tokensFormatados);
List <int> qtdRepeticaoTokens = dm.Calcular_frequencia_tokens(tokensStemmizados);
List <string> tokensRepetidosEliminados = fmtc.Retirar_itens_repetidos(tokensStemmizados);
treino tr = new treino(cst);
tr.Memorizar(tokensRepetidosEliminados, qtdRepeticaoTokens, emocao);
}
private static void DrummondsGenerate()
{
var stemmer = new Stemming();
string line;
var moods = new List <MoodItem>();
using (var reader = File.OpenText("wordlist.txt"))
{
string word;
string category;
while ((line = reader.ReadLine()) != null)
{
var split = line.Split(',');
if (split.Length >= 2)
{
double weight = 1;
if (split.Length > 2)
{
weight = double.Parse(split[2]);
}
moods.Add(new MoodItem
{
Word = stemmer.Stem(split[0].ToLower()),
Value = new MoodValue {
Category = split[1], Weight = weight
}
});
}
}
}
StreamWriter file2 = new StreamWriter("c:\\code\\files.txt", true);
foreach (var mood in moods)
{
file2.WriteLine("new MoodItem{ Word = \"" + mood.Word + "\", Value = new MoodValue { Category =\"" + mood.Value.Category + "\", Weight = " + mood.Value.Weight + " } },");
}
file2.Close();
}
public static string stem(string input)
{
string[] terms = input.Split(new char[] { ' ' }, StringSplitOptions.RemoveEmptyEntries);
List <string> list = new List <string>();
foreach (string term in terms)
{
Stemming ss = new Stemming();
string s = ss.Stem(term);
list.Add(s);
}
StringBuilder sb = new StringBuilder();
for (int i = 0; i < list.Count(); i++)
{
if (i != 0)
{
sb.Append(" ");
}
sb.Append(list[i]);
}
return(sb.ToString());
}
