//~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~//
/// <summary>
///
/// </summary>
/// <param name="Unit"></param>
/// <param name="LayerTotheRight"></param>
/// <returns></returns>
private double Sigmoud(Node Unit , ArrayList LayerTotheRight)
{
if (LayerTotheRight == null) //Neuron in Output Layer
return (Unit.Output * (Unit.Target - Unit.Output) * ( 1 - Unit.Output));
else //Neuron in Hidden Layer
return (Unit.Output * ( 1 - Unit.Output) * SumOF(LayerTotheRight , Unit.Number) );
}
//~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~//
/// <summary>
/// For each input vector presented,
/// the Euclidean distance to the output node is computed
/// </summary>
/// <param name="Input">input vector</param>
/// <param name="Unit">the output</param>
/// <returns>the Euclidean distance</returns>
protected double Distance(ArrayList Input , Node Unit)
{
double Euclidean_distance = 0,Xi;
for(int i = 0 ; i < Input.Count ; i++)
{
Xi = ((Node)Input[i]).Output;
Euclidean_distance += ((Xi - Unit[i]) * (Xi - Unit[i])) ;
}
return Euclidean_distance;
}
//~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~//
/// <summary>
///
/// </summary>
/// <param name="CurrentUnit"></param>
/// <param name="PreviousLayer"></param>
/// <returns></returns>
private double Net(Node CurrentUnit, ArrayList PreviousLayer )
{
double NET = 0;
foreach(Node Unit in PreviousLayer )
{
NET += Unit.Output * CurrentUnit.Weight(Unit.Number);
}
return (NET + CurrentUnit.Bias);
}
