public void LarningOnSingleInput(Matrix <double> input, double radius)
{
var bestIndex = _network.GetIndexOfTheMostMatchesNeuron(input);
var candidateToLearning = _network.GetCandidatesToLearning(bestIndex, radius);
candidateToLearning.ForEach(x => TeachCandidate(x, input, radius));
}
public override int Run(string[] remainingArguments)
{
InitializeArgumentsFromConfinIfItIsNecessary();
var inputs = _dataLoader.LoadCoordinates(_fileName);
var normalizedInputs = inputs.Select(_matrixHelpers.Normalize).ToList();
var maxInputOne = normalizedInputs.Select(x => x.At(0, 0)).Max();
var maxInputTwo = normalizedInputs.Select(x => x.At(0, 1)).Max();
var network = new SOMNet(maxInputOne, maxInputTwo, inputs.Count);
var rmax = 4;
//var neighborhoodFunction = new Guassian();
var neighborhoodFunction = new Guassian();
var teacher = new Teacher(network, neighborhoodFunction, rmax);
teacher.Learn(normalizedInputs);
var order = new List <int>();
foreach (var input in inputs)
{
order.Add(network.GetIndexOfTheMostMatchesNeuron(input));
}
var saver = new DataSaver();
saver.SaveWeights(network);
saver.SaveNormalizedInputs(normalizedInputs);
saver.SaveOrder(order);
if (shouldVisual)
{
var orderlinesses = inputs.Select((x, i) => new Orderliness(order[i], x)).ToList();
var bm = _visualizer.Visualize(orderlinesses);
bm.Save(ConfigurationManager.AppSettings["Visualization"]);
}
return(0);
}