private WinningNeuronModel winningNeuron(Position iteration, int counter, bool isDisabler, int IterationDisable, int CycleQty)
{
WinningNeuronModel winningNeuron = new WinningNeuronModel()
{
Value = null, Id = 0
};
neurons.ForEach(n =>
{
if (isDisabler && (CycleQty == 0 || counter < CycleQty) && n.StopCount <= IterationDisable && n.StopCount > 0)
{
n.StopCount++;
}
else
{
n.StopCount = 0;
var d = SimiliarityCalc(n.Position, iteration) + n.Balance;
if (winningNeuron.Value == null || d < winningNeuron.Value)
{
winningNeuron.Value = d;
winningNeuron.Id = n.Id;
}
}
});
return(winningNeuron);
}
public void Calc(double LearnRation, int balance = 0, bool isDisabler = false, int IterationDisable = 0, int CycleQty = 0)
{
int counter = 0;
dataList.ForEach(it => {
WinningNeuronModel winningNeuron = this.winningNeuron(it, counter, isDisabler, IterationDisable, CycleQty);
GetNewWeight(winningNeuron, it, LearnRation, balance);
counter++;
});
}
private void GetNewWeight(WinningNeuronModel winningNeuron, Position iteration, double LearnRation, int balance)
{
var position = neurons.Find(p => p.Id == winningNeuron.Id).Position;
position.X = (int)Math.Round(position.X + (LearnRation * (iteration.X - position.X)), 3);
position.Y = (int)Math.Round(position.Y + (LearnRation * (iteration.Y - position.Y)), 3);
neurons.Find(p => p.Id == winningNeuron.Id).Position = new Position()
{
X = position.X, Y = position.Y
};
neurons.Find(p => p.Id == winningNeuron.Id).Balance += balance;
neurons.Find(p => p.Id == winningNeuron.Id).StopCount = 1;
}
