/**
* Prints network output for the each element from the specified training set.
* @param neuralNet neural network
* @param trainingSet training set
*/
public static void TestNeuralNetwork(NeuralNetwork neuralNet, TrainingSet trainingSet)
{
foreach (TrainingElement trainingElement in trainingSet.TrainingElements)
{
neuralNet.SetInput(trainingElement.Input);
neuralNet.Calculate();
double[] networkOutput = neuralNet.Output;
Console.Write("Input: " + trainingElement.Input.ArrayString());
Console.WriteLine(" Output: " + networkOutput.ArrayString());
}
}
/// <summary>
/// Creates new PerceptronLearning learning for the specified neural network
/// </summary>
/// <param name="neuralNetwork"></param>
public PerceptronLearning(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
}
/// <summary>
/// Creates new instance of InstarLearning algorithm for the specified
/// neural network.
/// </summary>
/// <param name="neuralNetwork">neural network to train</param>
public InstarLearning(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
this.LearningRate = 0.1;
}
/// <summary>
/// Creates new SigmoidDeltaRule for the specified neural network
/// </summary>
/// <param name="neuralNetwork">neural network to train</param>
public SigmoidDeltaRule(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
}
/// <summary>
/// Sets default input and output neurons for network (first layer as input,
/// last as output)
/// </summary>
/// <param name="nnet"></param>
public static void SetDefaultIO(NeuralNetwork nnet)
{
IList<Neuron> inputNeurons = new List<Neuron>();
Layer firstLayer = (Layer)nnet.Layers[0];
foreach (Neuron neuron in firstLayer.Neurons)
{
if (!(neuron is BiasNeuron))
{ // dont set input to bias neurons
inputNeurons.Add(neuron);
}
}
IList<Neuron> outputNeurons = ((Layer)nnet.Layers[nnet.Layers.Count - 1]).Neurons;
nnet.InputNeurons = inputNeurons;
nnet.OutputNeurons = outputNeurons;
}
/**
* Predict sunspots.
* @param network Neural network to use.
*/
public void Predict(NeuralNetwork network)
{
Console.WriteLine("Year\tActual\tPredict\tClosed Loop Predict");
for (int year = EVALUATE_START; year < EVALUATE_END; year++)
{
// calculate based on actual data
double[] input = new double[WINDOW_SIZE];
for (int i = 0; i < input.Length; i++)
{
input[i] = this.normalizedSunspots[(year - WINDOW_SIZE) + i];
}
network.SetInput(input);
network.Calculate();
double[] output = network.Output;
double prediction = output[0];
this.closedLoopSunspots[year] = prediction;
// calculate "closed loop", based on predicted data
for (int i = 0; i < input.Length; i++)
{
input[i] = this.closedLoopSunspots[(year - WINDOW_SIZE) + i];
}
network.SetInput(input);
network.Calculate();
output = network.Output;
double closedLoopPrediction = output[0];
// display
Console.WriteLine((STARTING_YEAR + year) + "\t"
+ this.normalizedSunspots[year] + "\t"
+ prediction + "\t"
+ closedLoopPrediction);
}
}
public DynamicBackPropagation(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
}
/// <summary>
/// Creates new instance of CompetitiveLearning for the specified neural network
/// </summary>
/// <param name="neuralNetwork"></param>
public CompetitiveLearning(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
}
/// <summary>
/// Creates new HopfieldLearning for the specified neural network
/// </summary>
/// <param name="neuralNetwork"></param>
public HopfieldLearning(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
}
/// <summary>
/// Creates new instance of SupervisedHebbianLearning algorithm for the specified
/// neural network.
/// </summary>
/// <param name="neuralNetwork">neural network to train</param>
public SupervisedHebbianLearning(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
}
/// <summary>
/// Creates new instance of BinaryHebbianLearning for the specified neural network
/// </summary>
/// <param name="neuralNetwork"></param>
public BinaryHebbianLearning(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
}
/// <summary>
/// Creates new LMS learning rule for specified neural network
/// </summary>
/// <param name="neuralNetwork">neural network to train</param>
public LMS(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
}
/// <summary>
/// Creates an instance of OjaLearning algorithm for the specified
/// neural network
/// </summary>
/// <param name="neuralNetwork">neural network to train</param>
public OjaLearning(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
}
/// <summary>
/// Creates new BinaryDeltaRule learning for the specified neural network
/// </summary>
/// <param name="neuralNetwork"></param>
public BinaryDeltaRule(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
}
/// <summary>
/// Creates an instance of UnsupervisedHebbianLearning algorithm for the specified
/// neural network
/// </summary>
/// <param name="neuralNetwork">neural network to train</param>
public UnsupervisedHebbianLearning(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
this.LearningRate = 0.1;
}
/// <summary>
/// Creates new instance of MomentumBackpropagation learning for the specified neural network
/// </summary>
/// <param name="neuralNetwork">neural network to train</param>
public MomentumBackpropagation(NeuralNetwork neuralNetwork)
: base(neuralNetwork)
{
Momentum = 0.25d;
}
