private void CreateNetwork(NeuralLayer connectingFrom, NeuralLayer connectingTo)
{
foreach (var to in connectingTo.Neurons)
{
foreach (var from in connectingFrom.Neurons)
{
to.dendrites.Add(new Dendrite()
{
InputPulse = to.pulseOutput, SynapticWeight = connectingTo.Weight
});
}
}
}
public void AddLayer(NeuralLayer layer)
{
int denriteCount = 1;
if (layers.Count > 0)
{
denriteCount = layers[layers.Count - 1].Neurons.Count;
}
foreach (var element in layer.Neurons)
{
for (int i = 0; i < denriteCount; i++)
{
element.dendrites.Add(new Dendrite());
}
}
}
// Create a neural net, optionally specifying a seed for the Random object.
// This can be useful for using the same seed repeatedly to get repeated
//identical results.
public NeuralNetwork(int inputNeuronCount, int hiddenNeuronCount, int outputNeuronCount, double learningRate, int?randomSeed = null)
{
rand = randomSeed.HasValue ? new Random(randomSeed.Value) : new Random();
this.learningRate = learningRate;
InputLayer = new NeuralLayer();
HiddenLayer = new NeuralLayer();
OutputLayer = new NeuralLayer();
// Create the neurons for each layer
for (int i = 0; i < inputNeuronCount; i++)
{
InputLayer.Add(new Neuron(0)); // Input neurons do not have a bias
}
for (int i = 0; i < hiddenNeuronCount; i++)
{
HiddenLayer.Add(new Neuron(rand.NextDouble())); // A random bias, will be improved during training
}
for (int i = 0; i < outputNeuronCount; i++)
{
OutputLayer.Add(new Neuron(rand.NextDouble())); // A random bias, will be improved during training
}
// Wire up the input layer to the hidden layer
for (int i = 0; i < hiddenNeuronCount; i++)
{
for (int j = 0; j < inputNeuronCount; j++)
{
HiddenLayer[i].InputSignals.Add(InputLayer[j].OutputSignal, new NeuralFactor(CalcInitWeight(inputNeuronCount)));
}
}
// Wire up the hidden layer to the output layer
for (int i = 0; i < outputNeuronCount; i++)
{
for (int j = 0; j < hiddenNeuronCount; j++)
{
OutputLayer[i].InputSignals.Add(HiddenLayer[j].OutputSignal, new NeuralFactor(CalcInitWeight(hiddenNeuronCount)));
}
}
}
