public static void RecalculateWeights(NeuronForManipulation neuron, double eta, IList <double> inputs)
{
double etaDelta = eta * neuron.Delta; //improve performance
for (int i = 0; i < neuron.Weights.Count; i++)
{
neuron.Weights[i].Weight += etaDelta * inputs[i];
}
neuron.Bias += etaDelta;
}
public static void CalculateNeuronOutput(NeuronForManipulation neuron, ActivationFunctionType activationFunction, double Alpha, IList <double> inputs)
{
neuron.Weights = neuron.Weights.OrderBy(a => a.Index).ToList();
double outputSum = 0;
for (int i = 0; i < neuron.Weights.Count; i++)
{
outputSum += inputs[i] * neuron.Weights[i].Weight;
}
outputSum += neuron.Bias;
neuron.Output = NeuronActiveFunctionResult(activationFunction, Alpha, outputSum);
}
public static double NeuronDiffActiveFunctionResult(NeuronForManipulation neuron, ActivationFunctionType activationFunction, double Alpha)
{
var diffOutput = 1d;
switch (activationFunction)
{
case ActivationFunctionType.Lineal:
break;
case ActivationFunctionType.Tangential:
diffOutput = ((1d - neuron.Output) * (1d + neuron.Output));
break;
case ActivationFunctionType.Sigmoid:
diffOutput = Alpha * neuron.Output * (1d - neuron.Output);
break;
}
return(diffOutput);
}
public static void RecalculateDelta(NeuronForManipulation neuron, ActivationFunctionType activationFunction, double Alpha, double value)
{
//if hidden type, neuron receive a gradient, if output type, neuron receive an error.
neuron.Delta = value * NeuronDiffActiveFunctionResult(neuron, activationFunction, Alpha);
}
