public override List <double> Train(INetwork network, IDataSet data, IStatusReporter statusHolder)
{
vSetStart = (int)((1 - vSetPercentage) * data.PatternCount);
Normalizor.GetMinMaxActivationWithMargin(network.Activation.MinValue, network.Activation.MaxValue,
out minValidOutput, out maxValidOutput);
isRecursiveNet = network is RecursiveNetwork;
return(base.Train(network, data, statusHolder));
}
public void Normalize(double _maxValueFrom, double _minValueFrom, double _maxValueTo, double _minValueTo)
{
normalizor = new Normalizor(_minValueFrom, _maxValueFrom, _minValueTo, _maxValueTo);
for (int i = 0; i < cases.Count; i++)
{
GetInput(i).MapInplace(x => normalizor.Normalize(x));
if (HasOutput && ProblemType == ProblemType.Regression)
{
GetIdealOutput(i).MapInplace(x => normalizor.Normalize(x));
}
}
}
public void TestNormalizorMinMaxGetter()
{
double min;
double max;
double marginFactor = Normalizor.MARGIN_FACTOR;
Normalizor.GetMinMaxActivationWithMargin(Unipolar.MinValue, Unipolar.MaxValue, out min, out max);
double span = (Unipolar.MaxValue - Unipolar.MinValue);
double desiredMin = Unipolar.MinValue + marginFactor * span;
double desiredMax = Unipolar.MaxValue - marginFactor * span;
Assert.AreEqual(desiredMin, min);
Assert.AreEqual(desiredMax, max);
}
protected virtual double CalculateMSEError(INetwork network, IDataSet dataSet)
{
double mseSum = 0.0;
foreach (Pattern p in dataSet.EnumeratePatterns())
{
mseSum += MSECalculator.CalculateRawMSE(p.IdealOutput - network.ComputeOutput(p.Input));
}
double min;
double max;
Normalizor.GetMinMaxActivationWithMargin(network.Activation.MinValue, network.Activation.MaxValue, out min, out max);
return(MSECalculator.CalculateEpochMSE(mseSum, dataSet.PatternCount, min, max));
}
public void Update()
{
foreach (var v in data)
{
if (v < MinValue)
{
MinValue = v;
}
if (v > MaxValue)
{
MaxValue = v;
}
}
Normalizor = null;
}
