public void Train(string textToParse)
{
string[] textOriginal = textToParse.Split(' ');
List <string> textWithoutNumbers = new List <string>();
foreach (var word in textOriginal)
{
var removedNumberWord = Regex.Replace(word, @"\d", "");
textWithoutNumbers.Add(removedNumberWord.ToLower());
}
List <string> wordsList = new List <string>();
foreach (var word in textOriginal)
{
var removedPunctuationWord = Regex.Replace(word, @"[^\w\s]", ""); //\s is whitespace characters, namely spacebar and tab. \w is alphanumeric characters.
wordsList.Add(removedPunctuationWord.ToLower());
}
var commonWords = new CommonWords();
wordsList.RemoveAll(x => commonWords.commonWords.Contains(x));
wordsList.RemoveAll(x => x == String.Empty);
wordsList = wordsList.Distinct().ToList();
foreach (var str in wordsList)
{
if (wordCounter.ContainsKey(str.ToLower()))
{
wordCounter[str.ToLower()] += 1;
}
else
{
wordCounter.Add(str.ToLower(), 1);
}
}
}
private Tuple <string, double> CheckIfSpam(string text,
int countSpamMails, Dictionary <string, int> spamWordList,
int countNotSpamMails, Dictionary <string, int> notSpamWordList)
{
string[] textOriginal = text.Split(' ');
List <string> textWithoutNumbers = new List <string>();
foreach (var word in textOriginal)
{
var removedNumberWord = Regex.Replace(word, @"\d", "");
textWithoutNumbers.Add(removedNumberWord.ToLower());
}
List <string> wordsList = new List <string>();
foreach (var word in textWithoutNumbers)
{
var removedPunctuationWord = Regex.Replace(word, @"[^\w\s]", "");
wordsList.Add(removedPunctuationWord.ToLower());
}
var wordsToRemove = new CommonWords();
wordsList.RemoveAll(x => wordsToRemove.commonWords.Contains(x));
wordsList.RemoveAll(x => x == String.Empty);
wordsList = wordsList.Distinct().ToList();
List <double> PvaluesSpam = new List <double>();
List <double> PvaluesHam = new List <double>();
foreach (var word in wordsList)
{
if (notSpamWordList.ContainsKey(word) && spamWordList.ContainsKey(word))
{
var p = CalculateProbability(word.ToLower(),
countSpamMails, spamWordList,
countNotSpamMails, notSpamWordList);
PvaluesSpam.Add(p); //P(S|W)
PvaluesHam.Add(1 - p); //P(H|W)
}
}
//konacni izracun
double probabilityOfSpamMail;
if (PvaluesSpam.Count == 0 || PvaluesHam.Count == 0)
{
probabilityOfSpamMail = 0.5;// u slučaju da su nepoznate sve riječi emaila
}
else
{
double PSpam = PvaluesSpam.Aggregate((a, x) => a * x);
double PHam = PvaluesHam.Aggregate((a, x) => a * x);
probabilityOfSpamMail = PSpam / (PSpam + PHam);
}
if (probabilityOfSpamMail > 0.5)
{
return(Tuple.Create("SPAM", probabilityOfSpamMail));
}
else
{
return(Tuple.Create("HAM", probabilityOfSpamMail));
}
}
