static void netTest()
{
Random rand = new Random();
LearningDataSet dataSet = new LearningDataSet(2, 1);
for (int i = 0; i < 1000; i++)
{
double x = rand.NextDouble();
double y = rand.NextDouble();
if (rand.NextDouble() > .5)
{
x *= -1;
}
if (rand.NextDouble() > .5)
{
y *= -1;
}
double w = 1 / (1 + Math.Exp(-1 * (x * 1 + .25 * y)));
double z = 1 / (1 + Math.Exp(-1 * (x * .5 + 0 * y)));
double a = 1 / (1 + Math.Exp(-1 * (w - z)));
double b = 1 / (1 + Math.Exp(-1 * (w + z)));
double c = 1 / (1 + Math.Exp(-1 * (w)));
double d = 1 / (1 + Math.Exp(-1 * (z)));
dataSet.add(new TestInstance(new double[] { x, y }, new double[] { a, b, c, d }));
}
Console.WriteLine("Dataset loaded.");
NeuralNet net = new NeuralNet(new int[] { 2, 2, 4 });
Console.WriteLine(net.NodeCount());
Console.WriteLine(net.EdgeCount());
BackPropagationRunner bp = new BackPropagationRunner(dataSet, net);
bp.run(.5, 10000);
}
public BackPropagationRunner(LearningDataSet data, NeuralNet neuralNet)
{
this.data = data;
this.neuralNet = neuralNet;
}