private void xorBtn_Click(object sender, EventArgs e)
{
NeuralNet nn = new NeuralNet(2, 2, 1);
double[][] xor_examples = new double[][]
{
new double[] { 0.1, 0.1 }, // 0 0
new double[] { 0.1, 0.9 }, // 0 1
new double[] { 0.9, 0.1 }, // 1 0
new double[] { 0.9, 0.9 }, // 1 1
};
double[][] xor_targets = new double[][]
{
new double[] { 0.1 }, // 0
new double[] { 0.9 }, // 1
new double[] { 0.9 }, // 1
new double[] { 0.1 }, // 0
};
nn.train(xor_examples, xor_targets, Math.Pow(10, -20));
Debug.WriteLine("done learning");
Func <double, int> decodeOutput = x => {
if (x >= 0.5)
{
return(1);
}
return(0);
};
for (int j = 0; j < xor_examples.Length; j++)
{
double[] output = nn.run(xor_examples[j]);
for (int i = 0; i < output.Length; i++)
{
Debug.WriteLine(i + ": " + output[i] + " => " + decodeOutput(output[i]));
}
}
}
private void learnPctBtn_Click(object sender, EventArgs e)
{
NeuralNet nn = new NeuralNet(2, 6, 3);
double[][] examples = new double[pts.Length][];
for (int i = 0; i < pts.Length; i++)
{
examples[i] = new double[2];
examples[i][0] = pts[i].p.X;
examples[i][1] = pts[i].p.Y;
}
double[][] targets = new double[pts.Length][];
for (int i = 0; i < pts.Length; i++)
{
targets[i] = new double[zones.Length];
for (int j = 0; j < zones.Length; j++)
{
targets[i][j] = 0.1;
}
targets[i][pts[i].zone] = 0.9;
}
nn.train(examples, targets, 20);// Math.Pow(10, -1));
Debug.WriteLine("done learning");
Action <int, int, double[]> decodeOutput = (x, y, output) => {
double max = 0;
int pos = 0;
for (int i = 0; i < output.Length; i++)
{
if (output[i] > max)
{
max = output[i];
pos = i;
}
}
Point p = cartesianToScreenCoords(new Point(x, y));
((Bitmap)pictureBox1.Image).SetPixel(p.X, p.Y, zones[pos].pen.Color);
//Graphics g = pictureBox1.CreateGraphics();
//g.DrawEllipse(zones[pos].pen, p.X, p.Y, 3, 3);
//pictureBox1.Refresh();
};
for (int i = -WIDTH / 2; i < WIDTH / 2; i += 1)
{
for (int j = -WIDTH / 2; j < WIDTH / 2; j += 1)
{
double[] output = nn.run(new double[] { j, i });
decodeOutput(j, i, output);
}
}
pictureBox1.Refresh();
}