private void Init()
{
//Init classes
scalingFunction = new ScalingFunction();
activationFunctions = new ActivationFunctions();
geneticAlgorithm = new GeneticAlgorithm();
neuralNetwork = new NeuralNetwork();
weightsGenerator = new WeightsGenerator();
//Init Lists
keyStore = new List <BTCKeyStore>();
dataSet = new List <DataSet>();
valkeyStore = new List <BTCKeyStore>();
valdataSet = new List <DataSet>();
nnld = new List <NeuralNetworkLayerDesign>();
currentMaxBytes = 0;
deathRate = 10; //If too high, then chance plays an increasing role and skews the result.
}
private byte ExecuteNetwork(double[] exampleAddressDbl)
{
NeuralNetwork nn = new NeuralNetwork();
ScalingFunction sf = new ScalingFunction();
ActivationFunctions af = new ActivationFunctions();
double[][] networkWeights = new double[5][];
double[][] networkBias = new double[5][];
WeightsGenerator wg = new WeightsGenerator();
int[] networks1 = new int[] { 8, 6, 4, 2, 1 };
networkWeights[0] = wg.CreateWeights(3, networks1[0] * 32);
networkWeights[1] = wg.CreateWeights(3, networks1[1] * 8);
networkWeights[2] = wg.CreateWeights(3, networks1[2] * 6);
networkWeights[3] = wg.CreateWeights(3, networks1[3] * 4);
networkWeights[4] = wg.CreateWeights(3, networks1[4] * 2);
networkBias[0] = Enumerable.Repeat(0.00, networks1[0] * 32).ToArray();
networkBias[1] = Enumerable.Repeat(0.00, networks1[1] * 8).ToArray();
networkBias[2] = Enumerable.Repeat(0.00, networks1[2] * 6).ToArray();
networkBias[3] = Enumerable.Repeat(0.00, networks1[3] * 4).ToArray();
networkBias[4] = Enumerable.Repeat(0.00, networks1[4] * 2).ToArray();
double[] layerOneOutput = nn.PerceptronLayer(networks1[0], exampleAddressDbl, networkWeights[0], exampleAddressDbl.Length, networkBias[0]);
layerOneOutput = af.TanSigmoid(layerOneOutput);
double[] layerTwoOutput = nn.PerceptronLayer(networks1[1], layerOneOutput, networkWeights[1], layerOneOutput.Length, networkBias[1]);
layerTwoOutput = af.TanSigmoid(layerTwoOutput);
double[] layerThreeOutput = nn.PerceptronLayer(networks1[2], layerTwoOutput, networkWeights[2], layerTwoOutput.Length, networkBias[2]);
layerThreeOutput = af.TanSigmoid(layerThreeOutput);
double[] layerFourOutput = nn.PerceptronLayer(networks1[3], layerThreeOutput, networkWeights[3], layerThreeOutput.Length, networkBias[3]);
layerFourOutput = af.TanSigmoid(layerFourOutput);
double[] layerOutput = nn.PerceptronLayer(networks1[4], layerFourOutput, networkWeights[4], layerFourOutput.Length, networkBias[4]);
layerOutput = af.TanSigmoid(layerOutput);
return((byte)(255.0 * layerOutput[0]));
}
