// probability for just one file
private List <WordMetrics> CalculateProbability(string categoryName, List <string> uniqueVocabulary,
Dictionary <string, int> categoryFile, int numberOfTrainingFiles)
{
// number of unique words in the training set
int uniqueWords = uniqueVocabulary.Count();
int allCatWords = 0;
List <WordMetrics> l_wordMetrics = new List <WordMetrics>();
double categoryProbability = 1 / (double)numberOfTrainingFiles;
// calculate number of all the words in the category
foreach (KeyValuePair <string, int> word in categoryFile)
{
allCatWords += word.Value;
}
// create the first object for the category and its probability
WordMetrics categoryWord = new WordMetrics(categoryName, 1, categoryProbability, uniqueWords, allCatWords);
l_wordMetrics.Add(categoryWord);
// calculate probability for each word and then create a new object with this information
foreach (KeyValuePair <string, int> word in categoryFile)
{
double probabilityResult = ((word.Value + 1) / (double)(uniqueWords + allCatWords));
WordMetrics wordMetrics = new WordMetrics(word.Key, word.Value, probabilityResult, uniqueWords, allCatWords);
l_wordMetrics.Add(wordMetrics);
}
return(l_wordMetrics);
}
// probability for more than one file in the category
private List <WordMetrics> CalculateProbability(string categoryName, List <string> uniqueVocabulary,
List <Dictionary <string, int> > categoryFiles, int numberOfTrainingFiles)
{
// number of unique words in the training set
int uniqueWords = uniqueVocabulary.Count();
int allCatWords = 0;
Dictionary <string, int> dictCopy = new Dictionary <string, int>();
List <WordMetrics> l_wordMetrics = new List <WordMetrics>();
double categoryProbability = categoryFiles.Count() / (double)numberOfTrainingFiles;
// merge two or more dictionaries for the category together
// if the keys are the same just add their values
int frequencyValue = 0;
foreach (Dictionary <string, int> dictionary in categoryFiles)
{
foreach (KeyValuePair <string, int> kvp in dictionary)
{
if (dictCopy.ContainsKey(kvp.Key))
{
dictCopy[kvp.Key] += kvp.Value;
}
else
{
dictCopy.Add(kvp.Key, kvp.Value);
}
}
}
// calculate number of all the words in the category
foreach (KeyValuePair <string, int> frequency in dictCopy)
{
allCatWords += frequency.Value;
}
// creates the first object for the category and its probability
WordMetrics categoryWord = new WordMetrics(categoryName, categoryFiles.Count(), categoryProbability, uniqueWords, allCatWords);
l_wordMetrics.Add(categoryWord);
// calculate probability for each word and then create a new object with this information
foreach (KeyValuePair <string, int> word in dictCopy)
{
double probabilityResult = (word.Value + 1) / (double)(uniqueWords + allCatWords);
WordMetrics wordMetrics = new WordMetrics(word.Key, word.Value, probabilityResult, uniqueWords, allCatWords);
l_wordMetrics.Add(wordMetrics);
}
return(l_wordMetrics);
}