public async Task PerformTests()
{
await Task.Run(() =>
{
Console.WriteLine("Szykuje dane zbioru uczacego");
double[][] neuralLearningInput = NetworkHelper.CreateLearningInputDataSet(learningFaces);
double[][] neuralLearningOutput = NetworkHelper.CreateLearningOutputDataSet(learningFaces, 4);
double[][] neuralTestingInput = NetworkHelper.CreateLearningInputDataSet(testingFaces);
double[][] neuralTestingOutput = NetworkHelper.CreateLearningOutputDataSet(testingFaces, 4);
INeuralDataSet learningSet, testingSet;
learningSet = NetworkHelper.NormaliseDataSet(neuralLearningInput, neuralLearningOutput);
testingSet = NetworkHelper.NormaliseDataSet(neuralTestingInput, neuralTestingOutput);
TestsHelper.RunTests(learningSet, testingSet, learningFaces[0].Features.Count, testingFaces.Count, learningFaces, testingFaces);
});
}
public static List <Tuple <INeuralDataSet, INeuralDataSet> > GetCrossValidationData(List <Face> learningFaces, List <Face> testingFaces, int pairs)
{
int testSize = learningFaces.Count;
Random rnd = new Random();
List <Tuple <INeuralDataSet, INeuralDataSet> > crossValidation = new List <Tuple <INeuralDataSet, INeuralDataSet> >();
List <Face> allData = new List <Face>();
foreach (var f in learningFaces)
{
allData.Add(f);
}
foreach (var f in testingFaces)
{
allData.Add(f);
}
List <Face> tmpLearningFaces = new List <Face>();
List <Face> tmpTestingFaces = new List <Face>();
for (int i = 0; i < pairs; i++)
{
tmpLearningFaces = new List <Face>();
tmpTestingFaces = new List <Face>();
RandomizeSet(rnd, allData);
for (int j = 0; j < allData.Count; j++)
{
if (j < testSize)
{
tmpLearningFaces.Add(allData[j]);
}
else
{
tmpTestingFaces.Add(allData[j]);
}
}
double[][] neuralLearningInput = NetworkHelper.CreateLearningInputDataSet(tmpLearningFaces);
double[][] neuralLearningOutput = NetworkHelper.CreateLearningOutputDataSet(tmpLearningFaces, 4);
INeuralDataSet learningSet = NetworkHelper.NormaliseDataSet(neuralLearningInput, neuralLearningOutput);
double[][] neuralTestingInput = NetworkHelper.CreateLearningInputDataSet(tmpTestingFaces);
double[][] neuralTestingOutput = NetworkHelper.CreateLearningOutputDataSet(tmpTestingFaces, 4);
INeuralDataSet testingSet = NetworkHelper.NormaliseDataSet(neuralTestingInput, neuralTestingOutput);
crossValidation.Add(new Tuple <INeuralDataSet, INeuralDataSet>(learningSet, testingSet));
}
return(crossValidation);
}
public async Task PerformCalculation()
{
await Task.Run(() =>
{
Console.WriteLine("Szykuje dane zbioru uczacego");
double[][] neuralLearningInput = NetworkHelper.CreateLearningInputDataSet(learningFaces);
double[][] neuralLearningOutput = NetworkHelper.CreateLearningOutputDataSet(learningFaces, 4);
double[][] neuralTestingInput = NetworkHelper.CreateLearningInputDataSet(testingFaces);
double[][] neuralTestingOutput = NetworkHelper.CreateLearningOutputDataSet(testingFaces, 4);
INeuralDataSet learningSet, testingSet;
learningSet = NetworkHelper.NormaliseDataSet(neuralLearningInput, neuralLearningOutput);
testingSet = NetworkHelper.NormaliseDataSet(neuralTestingInput, neuralTestingOutput);
ITrain network = NetworkHelper.LearnNetwork(learningSet, testingSet, learningFaces[0].Features.Count, inputData, testingFaces.Count);
learnedNetwork = network;
});
}