public void ExtendDatabaseWithArticle(LabeledArticle article)
{
Annotation annotation = _processor.Annotate(article.Article);
HashSet <string> persons = NLPCoreHelper.GetPersons(annotation);
HashSet <string> organizations = NLPCoreHelper.GetOrganizations(annotation);
HashSet <string> locations = NLPCoreHelper.GetLocation(annotation);
HashSet <string> sportSpecificWords = NLPCoreHelper.GetOccurence(annotation, _sportSpecificWords);
foreach (string person in persons)
{
PersonOccurenceDatabase.AddWord(person, article.Category);
}
foreach (string organization in organizations)
{
OrganizationOccurenceDatabase.AddWord(organization, article.Category);
}
foreach (string location in locations)
{
LocationsOccurenceDatabase.AddWord(location, article.Category);
}
foreach (string sportSpecificWord in sportSpecificWords)
{
SportSpecificWordsOccurenceDatabase.AddWord(sportSpecificWord, article.Category);
}
}
public double TrainClassifierWithArticle(LabeledArticle article)
{
double[] input = GetRawFeatures(article.Article);
double[] output = GetOutput(article);
double error = _teacher.Run(input, output);
Console.WriteLine(error);
return(error);
}
public double CalculateError(LabeledArticle article)
{
double[] input = GetRawFeatures(article.Article);
double[] output = _classifier.Compute(input);
double sum = output.Sum();
for (int i = 0; i < output.Length; i++)
{
output[i] /= sum;
}
double[] correct = GetOutput(article);
double error = 0;
for (int i = 0; i < output.Length; i++)
{
error += Math.Abs(output[i] - correct[i]);
}
return(error);
}
private double[] GetOutput(LabeledArticle article)
{
double[] output = new double[_categoriesCount];
output[_categoryIndex[article.Category]] = 1;
return(output);
}
