public void execute()
{
double sum = 0.0;
double output = 0.0;
object[] fromLinks = this.getFromLinks().ToArray();
for (int inputLinksCount = 0; inputLinksCount < fromLinks.Length; inputLinksCount++)
{
sum += ((Link)fromLinks[inputLinksCount]).getWeight() * ((Link)fromLinks[inputLinksCount]).getInputValue();
}
sum = sum + this.bias;
if (transferFunction == 1)
{
output = (1 + Math.Exp(-sum));
output = 1 / output;
}
if (transferFunction == 2)
{
//n = 2/(1+exp(-2*n))-1 TANSIG
output = 2 / (1 + Math.Exp(-2 * sum)) - 1;
}
object[] toLinks = this.getToLinks().ToArray();
for (int outputLinksCount = 0; outputLinksCount < toLinks.Length; outputLinksCount++)
{
((Link)toLinks[outputLinksCount]).setInputValue(output);
}
if (this.GetType().ToString() == "Neurotic.NeuralNet+OutputNeuron")
{
NeuralNet.OutputNeuron n = (NeuralNet.OutputNeuron) this;
n.setOutputValue(output);
}
}
/// <summary>
/// The main entry point for the application.
/// </summary>
public float[] getNNOutput(int denormType) // denormType 0 (NO Denorm) OR 1(0~1) OR 2 (-1~1)
{
int numOutputs = net.getLayer(net.layers.Count - 1).getNeuronCount();
float[] f = new float[numOutputs];
for (int count = 0; count < numOutputs; count++)
{
NeuralNet.Layer layer = net.getLayer(net.layers.Count - 1);
NeuralNet.OutputNeuron o = (NeuralNet.OutputNeuron)layer.getNeuron(count);
f[count] = (float)Convert.ToDecimal((o.getOutputValues())[0]);
}
if (denormType == 0)
{
return(f);
}
else if (denormType == 1)
{
return(Denormalize(f, true));
}
else
{
return(Denormalize(f, false));
}
}
