static void Train_predict_and_save(string filepath)
{
// train & testing data
var training_data = new Dictionary <int, Tuple <float[], float[]> >
{ // ys xs
[0] = new Tuple <float[], float[]>(new float[] { 0, 1 }, new float[] { 1 }),
[1] = new Tuple <float[], float[]>(new float[] { 1, 0 }, new float[] { 1 }),
[2] = new Tuple <float[], float[]>(new float[] { 0, 0 }, new float[] { 0 }),
[3] = new Tuple <float[], float[]>(new float[] { 1, 1 }, new float[] { 0 })
};
var snn = new SimpleNeuralNetwork(2, 0.4f, Activation.FunctionsEnum.Sigmoid);
snn.Add(4);
snn.Add(1);
// train
Console.WriteLine("Entrenamiento:\n");
Random random = new Random();
int j = 0;
Console.WriteLine("Training ...");
for (int i = 0; i < 100000; i++)
{
j = random.Next(4);
snn.Train(training_data[j].Item1, training_data[j].Item2);
}
// predict
Console.WriteLine("\nPredicciones:\n");
for (int i = 0; i < 4; i++)
{
var res = snn.Predict(training_data[i].Item1);
Console.WriteLine(string.Format("xs [ {0}, {1} ] = {2}", training_data[i].Item1[0], training_data[i].Item1[1], res[0]));
}
SimpleNeuralNetwork.Save(snn, filepath);
Console.WriteLine("\nRed Neuronal guardada !!.\n");
}