/// <summary>
/// Goes through each record in the training set and displays the ideal output and the actual output.
/// </summary>
/// <param name="net">RBF network.</param>
/// <param name="trainingData"></param>
/// <param name="specieNames"></param>
public void QueryEquilateral(GreedyRandom gr, Network net, List <Record> trainingData, Dictionary <string, int> specieNames)
{
// invert the specie list
Dictionary <int, string> invSpecies = new Dictionary <int, string>();
foreach (string key in specieNames.Keys)
{
int value = specieNames[key];
invSpecies[value] = key;
}
Equilateral eq = new Equilateral(specieNames.Count); // create encoding matrix for specie names
// for each training record
foreach (Record rec in trainingData)
{
double[] output = net.ComputeRegression(rec.Input); // run the record through the RBF network
int idealIndex = eq.Decode(rec.Ideals); // find the ideal specie name for the record
int actualIndex = eq.Decode(output); // find the calculated specie name for the record
Console.WriteLine("Guess: " + invSpecies[actualIndex]
+ ", Ideal: " + invSpecies[idealIndex]);
}
Console.WriteLine("Final error: " + gr.LastError);
}
/// <summary>
/// Encodes each header using equilateral encoding.
/// </summary>
/// <param name="col">Column to encode.</param>
/// <returns>Returns dictionary containing specie name and index.</returns>
public Dictionary <string, int> EncodeEquilateral(int col)
{
string name = _headers[col]; // header name
Dictionary <string, int> classes = EnumerateClasses(col); // specie name and index
int classCount = classes.Count; // number of species
InsertColumns(col + 1, classCount - 1); // insert additional columns into _data
Equilateral eq = new Equilateral(classCount); // create equilateral for number of classes
// fill _data with encoded species
foreach (object[] obj in _data)
{
int idx = classes[obj[col].ToString()]; // get value for corresponding key
double[] encoded = eq.Encode(idx); // encode specie name
// fill _data with encoded values
for (int i = 0; i < classCount - 1; i++)
{
obj[col + i] = encoded[i];
}
}
// fill _headers with number of classes
for (int i = 0; i < classes.Count; i++)
{
_headers[col + i] = name + "-" + i;
}
return(classes); // return
}