private void Process()
{
bool isSuccess = false;
while (!isSuccess)
{
int max = FuzzyLogicUtil.FindMax(ChoiceValues.ToArray());
double normOfSimilarity = 0;
if (max != -1)
{
normOfSimilarity = F2.Neurons(max).Output();
}
else
{
// Indicate that there is no eligible neuron
// as winner, so create a dummy neuron in F2 layer.
max = F2.AddNeuron(FuzzyLogicUtil.MakeDummyNeuron(InputSize));
ChoiceValues.Add(
FuzzyLogicUtil.ChoicingValue(normOfInputs, 1.0 * InputSize, Choice));
normOfSimilarity = normOfInputs;
}
// Calculate and compare with vigilance
winningVigilance = normOfSimilarity / normOfInputs;
if (winningVigilance >= Vigilance)
{
winningNeuronPos = max;
isSuccess = true;
}
else
{
ChoiceValues[max] = 0;
}
}
// Adjust neuron weight automatically. if it is set.
if (AutoAdjustWeight)
{
AdjustWeight();
}
// Calculate output.
output = FuzzyLogicUtil.Intersect(F2.Neurons(WinningNeuronPos).Weights.ToArray(),
InputData, InputSize);
}
public void Run(double[] data)
{
input = data;
// Compute norm of input values.
normOfInputs = FuzzyLogicUtil.Norm(data);
// Transfer input values.
f1.Input(data);
// Calculate initial choicing value
ChoiceValues.Clear();
for (int i = 0; i < f2.Count; ++i)
{
double normOfWeight = FuzzyLogicUtil.Norm(F2.Neurons(i).Weights.ToArray());
ChoiceValues.Add(FuzzyLogicUtil.ChoicingValue(F2.Neurons(i).Output(), normOfWeight, Choice));
}
Process();
}
