static void Main(string[] args)
{
//instantiate neural net inputs and outputs
Net theNet = new Net();
Neuron outO = new Neuron();
theNet.AddOutput(outO);
Neuron outX = new Neuron();
theNet.AddOutput(outX);
for (int i = 0; i < 64; i++)
{
theNet.AddInput(new Neuron());
}
//BEGIN TRAINING
List <TrainingData> data = new List <TrainingData>();
//input training files
string fileName;
for (int i = 1; i < 56; i++)
{
fileName = "train" + i + ".txt";
string inText = System.IO.File.ReadAllText(fileName);
data.Add(new TrainingData(inText));
}
//train net
theNet.Train(data.ToArray());
//read in test files, convert to doubles, map to inputs
for (int i = 0; i < 8; i++)
{
fileName = "Test" + i + ".txt";
string inText = System.IO.File.ReadAllText(fileName);
string[] result = inText.Split(new string[] { ",", " ", "\n", "\r\n" }, StringSplitOptions.RemoveEmptyEntries);
for (int j = 1; j < 65; j++)
{
double temp = Convert.ToDouble(result[j]);
theNet.inputs[j - 1].Value = temp;
}
//evaluate the Net
theNet.Evaluate();
//output results
System.Console.WriteLine("OutO = " + outO.Value);
System.Console.WriteLine("OutX = " + outX.Value);
System.Console.WriteLine("\r\n");
}
//send synapse weights to output file
if (File.Exists("outputFile.txt"))
{
File.Delete("outputFile.txt");
}
FileStream outstream = new FileStream("outputFile.txt", FileMode.OpenOrCreate, FileAccess.Write);
StreamWriter writer = new StreamWriter(outstream);
string[] weights = theNet.getWeights();
for (int i = 0; i < 128; i = i + 16)
{
for (int j = i; j < (i + 16); j++)
{
writer.Write(weights[j] + ", ");
}
writer.Write("\r\n");
}
writer.Close();
//System.IO.File.WriteAllLines("outputFile.txt", theNet.synapses.Select(s => (s.Weight).ToString()));
System.Console.WriteLine("Press any key to exit.");
Console.ReadLine();
}
