static void Main(string[] args)
{
Console.WriteLine("Demonstrating functionality: hardcoded XOR problem:");
TestHardcodedXOR();
Console.WriteLine("Demonstrating functionality: simple network exercise");
TestWithSingleInput();
Console.WriteLine("\nPress [ENTER] to execute actual NEAT.");
Console.ReadLine();
string[] inputColumns = new string[]
{
"genres",
"director_name",
"budget",
};
string[] expectedOutputColumns = new string[]
{
"gross",
};
NormalizedDataAndDictionaries inputData = GetDataFromDatabase(inputColumns);
NormalizedDataAndDictionaries outputData = GetDataFromDatabase(expectedOutputColumns);
RunNeat(inputData.NormalizedData, outputData.NormalizedData, outputData.denormalizationVariablesList[0]);
Console.WriteLine("\nPress [ENTER] to terminate application.");
Console.ReadLine();
}
private void ExecuteNeat()
{
bool wait = false;
int inputCount = 2;
int outputCount = 1;
int delay = 150;
Dispatcher.Invoke(() =>
{
controlsToDisable.ForEach(x => x.IsEnabled = false);
controlsToEnable.ForEach(x => x.IsEnabled = true);
if (CbAnimate.IsChecked == true)
{
wait = true;
}
inputCount = int.TryParse(TbInputs.Text, out inputCount) ? inputCount : 2;
outputCount = int.TryParse(TbOutputs.Text, out outputCount) ? outputCount : 1;
delay = int.TryParse(TbDelay.Text, out delay) ? delay : 0;
});
Neat neat = new Neat();
neat.OnGenerationEnd += (a) =>
{
Dispatcher.Invoke(() =>
{
TBSynapseData.Clear();
a.synapses.ForEach(x =>
{
TBSynapseData.Text += "FromLayer: " + x.FromLayer + "\n";
TBSynapseData.Text += "Fromneuron: " + x.FromNeuron + "\n";
TBSynapseData.Text += "ToLayer: " + x.ToLayer + "\n";
TBSynapseData.Text += "ToNeuron: " + x.ToNeuron + "\n";
TBSynapseData.Text += "Weight: " + x.Weight + "\n";
TBSynapseData.Text += "\n";
});
TBNeuronData.Clear();
a.hiddenNeurons.ForEach(x =>
{
TBNeuronData.Text += "Layer: " + x.Layer + "\n";
TBNeuronData.Text += "NeuronPosition: " + x.NeuronPosition + "\n";
TBNeuronData.Text += "Bias: " + x.Bias + "\n";
TBNeuronData.Text += "\n";
});
TBOutputNeurons.Clear();
a.outputNeurons.ToList().ForEach(x =>
{
TBOutputNeurons.Text += "Value: " + x.Value + "\n";
TBOutputNeurons.Text += "\n";
});
});
if (wait)
{
DrawAnn(a);
Thread.Sleep(delay);
}
while (pause)
{
}
neat.LowerWaitFlag();
};
string[] inputColumns = new string[]
{
//"movie_title",
"genres",
"director_name",
"budget",
};
string[] expectedOutputColumns = new string[]
{
"gross",
};
NormalizedDataAndDictionaries inputData = GetDataFromDatabase(inputColumns);
NormalizedDataAndDictionaries outputData = GetDataFromDatabase(expectedOutputColumns);
//Ann ann = neat.Train(10000, 100, 0.1, 3, 30, wait, inputs, expectedOutputs, Crossover.TwoPointCrossover, ActivationFunction.Identity);
//Ann ann = neat.Train(1000, 100, 0.03, 3, 30, false, inputs, expectedOutputs, Crossover.TwoPointCrossover, ActivationFunction.Identity);
//Ann ann = neat.Train(1000, 100, 0.03, 3, 30, false, inputData.NormalizedData, outputData.NormalizedData, Crossover.BestParentClone, ActivationFunction.Sigmoid);
Ann ann = neat.Train(1000, 100, 0.03, 3, 30, false, inputData.NormalizedData, outputData.NormalizedData, Crossover.BestParentClone, ActivationFunction.Identity);
Dispatcher.Invoke(() =>
{
controlsToDisable.ForEach(x => x.IsEnabled = true);
controlsToEnable.ForEach(x => x.IsEnabled = false);
SVCanvas.Background = new SolidColorBrush(Colors.Gray);
});
if (!wait)
{
DrawAnn(ann);
}
}
