public TrainModel(FeatureExtractionType eType, int EuclideanCentroidPointIndex, ClassifierType cType)
{
this.cType = cType;
Trained = false;
if (eType == FeatureExtractionType.EuclideanDistance)
featureExtraction = new EuclideanDistance(EuclideanCentroidPointIndex);
}
public TrainModel(FeatureExtractionType eType, ClassifierType cType)
{
this.cType = cType;
Trained = false;
if (eType == FeatureExtractionType.EuclideanDistance)
featureExtraction = new EuclideanDistance(14);
}
/// <summary>
/// Extracts features from the doc specified.
/// </summary>
/// <param name="doc">The document.</param>
/// <param name="docs">The docs to use to compare.</param>
/// <param name="featureCount">The number of features/terms to return.</param>
/// <param name="language">The language/extraction algorithm to use.</param>
/// <returns>The important features/terms of the doc.</returns>
public string[] Extract(Document doc, Document[] docs, int featureCount, FeatureExtractionType language)
{
if (!Extractors.TryGetValue(language, out var Extractor))
{
return(Array.Empty <string>());
}
return(Extractor.Extract(doc, docs, featureCount));
}
public TrainModel(FeatureExtractionType eType, int EuclideanCentroidPointIndex, ClassifierType cType)
{
this.cType = cType;
Trained = false;
if (eType == FeatureExtractionType.EuclideanDistance)
{
featureExtraction = new EuclideanDistance(EuclideanCentroidPointIndex);
}
}
public TrainModel(FeatureExtractionType eType, ClassifierType cType)
{
this.cType = cType;
Trained = false;
if (eType == FeatureExtractionType.EuclideanDistance)
{
featureExtraction = new EuclideanDistance(14);
}
}
public KFoldCrossValidation(List <Data> data, int k_fold, ClassifierType cType, FeatureExtractionType fType, int K = 3)
{
Kfold = k_fold;
this.K = K;
this.cType = cType;
this.fType = fType;
Data = data;
r = new Random();
chunckSize = new List <int>(Kfold);
for (int i = 0; i < Kfold - 1; ++i)
{
chunckSize.Add(Data.Count / Kfold);
}
chunckSize.Add(Data.Count - ((Kfold - 1) * (Data.Count / Kfold)));
}
/// <summary>
/// Gets the topics important in this document
/// </summary>
/// <param name="comparisonDocuments">The comparison documents.</param>
/// <param name="featureCount">The feature count.</param>
/// <param name="featureExtractionType">Type of the feature extraction.</param>
/// <returns>The features/topics important in this document.</returns>
public string[] GetFeatures(Document[] comparisonDocuments, int featureCount, FeatureExtractionType featureExtractionType)
{
return(FeatureExtractor.Extract(this, comparisonDocuments, featureCount, featureExtractionType));
}