public void Create(NeuralNetworkCls parent, int[] nrOfNeuronsList, int index, Synapses prevSy)
{
if (parent.CompleteObjList)
{
parent.NeuronsAndSynappses.Add(this);
}
Parent = parent;
PrevSy = prevSy; // could be null
int nrOfNeurons = nrOfNeuronsList[index];
I = Matrix <double> .Build.Dense(nrOfNeurons, 1);
O = Matrix <double> .Build.Dense(nrOfNeurons, 1);
if (parent.FunctionInitializer != null)
{
ResultType = parent.FunctionInitializer.GetResultEnum(index);
}
else
{
ResultType = (ResultEnum)index;
}
if (index < nrOfNeuronsList.Length - 1)
{
NextSy = new Synapses();
NextSy.Create(Parent, nrOfNeuronsList, index, this);
}
else
{
parent.LastNeurons = this;
}
}
private void CheckIfNeuralNetworksAreEqual(NeuralNetworkCls nnFromEcel, NeuralNetworkCls neuralNetworkCls)
{
for (int i = 0; i < nnFromEcel.NeuronsAndSynappses.Count; i++)
{
if (i % 2 == 1)
{
Synapses s1 = nnFromEcel.NeuronsAndSynappses[i] as Synapses;
Synapses s2 = neuralNetworkCls.NeuronsAndSynappses[i] as Synapses;
if (!s1.W.Equals(s2.W) || !s1.B.Equals(s2.B))
{
throw new Exception("W or B are not equal.");
}
}
}
//System.Windows.Forms.MessageBox.Show("Excel test completed. The current neural network delivered the same results as the serialised class.", "Success", System.Windows.Forms.MessageBoxButtons.OK, System.Windows.Forms.MessageBoxIcon.Asterisk);
}
