static void Evaluate(Perceptron p, double from, double to, double step)
{
string output = "";
for (double i = from; i < to; i += step)
{
double res = p.Activate(new double[] { normalize(i, inputMin, inputMax) })[0];
output += i + ";" + inverseNormalize(res, outputMin, outputMax) + "\n";
Console.WriteLine(i + ";" + res + "\n");
}
System.IO.File.WriteAllText(outputPath, output);
}
private static void Evaluate(Perceptron p, double from, double to, double step)
{
var output = "";
for (var i = from; i < to; i += step)
{
var res = p.Activate(new double[] { Normalize(i, inputMin, inputMax) })[0];
output += i + ";" + InverseNormalize(res, outputMin, outputMax) + "\n";
Console.WriteLine(i + ";" + res + "\n");
}
File.WriteAllText(outputPath, output);
}
static void Main(string[] args)
{
Perceptron p;
int[] net_def = new int[] { inputCount, 10, 10, outputCount };
double learning_rate = 0.3;
double max_error = 0.0001;
int max_iter = 1000000;
if (!loadNetwork)
{
ReadData();
p = new Perceptron(net_def);
while (!p.Learn(input, output, learning_rate, max_error, max_iter, neuralNetworkPath, 10000))
{
p = new Perceptron(net_def);
}
}
else
{
p = Perceptron.Load(neuralNetworkPath);
}
//Evaluate(p, 0, 5, 0.1);
while (true)
{
double[] val = new double[inputCount];
for (int i = 0; i < inputCount; i++)
{
Console.WriteLine("Inserte valor " + i + ": ");
val[i] = normalize(double.Parse(Console.ReadLine()), inputMin, inputMax);
}
double[] sal = p.Activate(val);
for (int i = 0; i < outputCount; i++)
{
Console.Write("Respuesta " + i + ": " + inverseNormalize(sal[i], outputMin, outputMax) + " ");
}
Console.WriteLine("");
}
}
public static Perceptron LoadNetwork(string neuralNetworkPath)
{
Perceptron p = null;
try
{
if (File.Exists(neuralNetworkPath))
{
var json = File.ReadAllText(neuralNetworkPath);
p = JsonConvert.DeserializeObject <Perceptron>(json);
}
}
catch (SerializationException ex)
{
Console.WriteLine("Failed to deserialize. Reason: " + ex.Message);
}
return(p);
//var fs = new FileStream(neuralNetworkPath, FileMode.Open);
//Perceptron p = null;
//try
//{
// var formatter = new BinaryFormatter();
// // Deserialize the hashtable from the file and
// // assign the reference to the local variable.
// p = (Perceptron)formatter.Deserialize(fs);
//}
//catch (SerializationException e)
//{
// Console.WriteLine("Failed to deserialize. Reason: " + e.Message);
// throw;
//}
//finally
//{
// fs.Close();
//}
//return p;
}
public static void Main()
{
Perceptron p;
var net_def = new int[] { inputCount, 10, 10, outputCount };
var learning_rate = 0.3;
var max_error = 0.0001;
var max_iter = 1000000;
loadNetwork = File.Exists(neuralNetworkPath);
if (!loadNetwork)
{
ReadData();
p = new Perceptron(net_def);
while (!p.Learn(input, output, learning_rate, max_error, max_iter, neuralNetworkPath, 10000))
{
p = new Perceptron(net_def);
}
}
else
{
p = Perceptron.LoadNetwork(neuralNetworkPath);
}
p.SaveNetwork(neuralNetworkPath);
//Evaluate(p, 0, 5, 0.1);
while (true)
{
var inputs = new double[inputCount];
for (var i = 0; i < inputCount;)
{
Console.WriteLine($"Value #{i}: ");
var inputValue = Console.ReadLine();
if (!string.IsNullOrEmpty(inputValue))
{
if (double.TryParse(inputValue, out var input))
{
var inputNormalized = Normalize(input, inputMin, inputMax);
Console.WriteLine($"{input:F3} => {inputNormalized:F5}");
inputs[i] = inputNormalized;
i++;
}
}
else
{
goto byebye;
}
}
var outputs = p.Activate(inputs);
//for (var i = 0; i < outputCount; i++)
//{
var sbInputs = new StringBuilder();
var sbOutputs = new StringBuilder();
foreach (var inp in inputs)
{
if (sbInputs.Length > 0)
{
sbInputs.Append(", ");
}
sbInputs.Append($"{inp:F5}");
}
foreach (var o in outputs)
{
var outValueDenormalized = InverseNormalize(o, outputMin, outputMax);
if (sbOutputs.Length > 0)
{
sbOutputs.Append(", ");
}
sbOutputs.Append($"{o:F3}");
}
Console.Write($"Inputs: {sbInputs} => Outputs: {sbOutputs}");
//}
Console.WriteLine("");
}
byebye:
Console.WriteLine("bye bye!");
}
static void Main(string[] args)
{
Perceptron p;
if (!loadNetwork)
{
ReadData();
p = new Perceptron(new int[] { inputCount, 5, 5, outputCount });
while (!p.Learn(input, output, 0.05, 0.01, 3000000))
{
p = new Perceptron(new int[] { inputCount, 5, 5, outputCount });
}
if (saveNetwork)
{
FileStream fs = new FileStream(neuralNetworkPath, FileMode.Create);
BinaryFormatter formatter = new BinaryFormatter();
try
{
formatter.Serialize(fs, p);
}
catch (SerializationException e)
{
Console.WriteLine("Failed to serialize. Reason: " + e.Message);
throw;
}
finally
{
fs.Close();
}
}
}
else
{
FileStream fs = new FileStream(neuralNetworkPath, FileMode.Open);
try
{
BinaryFormatter formatter = new BinaryFormatter();
// Deserialize the hashtable from the file and
// assign the reference to the local variable.
p = (Perceptron)formatter.Deserialize(fs);
}
catch (SerializationException e)
{
Console.WriteLine("Failed to deserialize. Reason: " + e.Message);
throw;
}
finally
{
fs.Close();
}
}
//Evaluate(p, 0, 5, 0.1);
while (true)
{
double[] val = new double[inputCount];
for (int i = 0; i < inputCount; i++)
{
Console.WriteLine("Inserte valor " + i + ": ");
val[i] = normalize(double.Parse(Console.ReadLine()), inputMin, inputMax);
}
double[] sal = p.Activate(val);
for (int i = 0; i < outputCount; i++)
{
Console.Write("Respuesta " + i + ": " + inverseNormalize(sal[i], outputMin, outputMax) + " ");
}
Console.WriteLine("");
}
}