/// <summary>
/// Merges base dimension of each weight dictionary
/// </summary>
/// <param name="results">The results.</param>
/// <returns></returns>
public static WeightDictionary MergeDimensions(this List <WeightDictionary> results)
{
WeightDictionary output = new WeightDictionary();
output.nDimensions = results.Count;
Dictionary <String, Double[]> tempMatrix = new Dictionary <string, double[]>();
Int32 id = 0;
foreach (var dict in results)
{
output.name += dict.name;
foreach (var en in dict.index)
{
if (!tempMatrix.ContainsKey(en.Key))
{
tempMatrix.Add(en.Key, new Double[output.nDimensions]);
}
tempMatrix[en.Key][id] = en.Value.weight;
}
id++;
}
foreach (var pair in tempMatrix)
{
output.AddEntry(pair.Key, pair.Value, false);
}
return(output);
}
/// <summary>
/// Constructs the weight dictionary according to stored frequency and score information
/// </summary>
/// <returns></returns>
public WeightDictionary ConstructWeightDictionary()
{
WeightDictionary output = new WeightDictionary();
var tkns = frequencyIndex.GetTokens();
Int32 maxFrequency = frequencyIndex.GetMaxFrequency();
Double maxWeight = scoreIndex.Values.Max();
Int32 maxDF = documentIndex.GetMaxFrequency();
foreach (String token in tkns)
{
Double finalWeight = scoreIndex[token].GetRatio(maxWeight); //.GetRatio(frequencyIndex.GetTokenFrequency(token));
Double TF = frequencyIndex.GetTokenFrequency(token).GetRatio(maxFrequency);
Double IDF = Math.Log(maxDF / documentIndex.GetTokenFrequency(token)) + 1; // Math.Log(1 - ( / maxDF));
finalWeight = finalWeight * (TF * IDF);
output.AddEntry(token, finalWeight);
}
return(output);
}