public override string ToString()
{
if (count < 1)
{
return(name + ": <no data>");
}
if (times)
{
double total_sec = total * 1.0 / Stopwatch.Frequency;
if (count == 1)
{
return(name + ": " + total_sec.ToString());
}
double average_sec = total_sec / count;
double sumsq_sec = sumsq / Stopwatch.Frequency / Stopwatch.Frequency;
double variance = (sumsq_sec - total_sec * total_sec / count) / (count - 1);
double stdev = Math.Sqrt(variance);
double conf95 = StudentT.InvCDF(0, stdev, count - 1, 0.975) / Math.Sqrt(count);
return(string.Format("Aggregate {0}: avg_usec {1} conf95plusorminus {2} stdev {3} count {4} sum {5}", name, average_sec * Math.Pow(10, 6), conf95 * Math.Pow(10, 6), stdev * Math.Pow(10, 6), count, total_sec * Math.Pow(10, 6)));
}
else
{
if (count == 1)
{
return(total.ToString());
}
double average = total * 1.0 / count;
double variance = (sumsq - (total * 1.0 * total) / count) / (count - 1);
double stdev = Math.Sqrt(variance);
double conf95 = StudentT.InvCDF(0, stdev, count - 1, 0.975) / Math.Sqrt(count);
return(string.Format("Aggregate {0}: avg_usec {1} conf95plusorminus {2} stdev {3} count {4} sum {5}", name, average * Math.Pow(10, 6), conf95 * Math.Pow(10, 6), stdev * Math.Pow(10, 6), count, total * Math.Pow(10, 6)));
}
}
public void ValidateInverseCumulativeDistribution(double location, double scale, double dof, double x, double p)
{
var dist = new StudentT(location, scale, dof);
Assert.That(dist.InverseCumulativeDistribution(p), Is.EqualTo(x).Within(1e-6));
Assert.That(StudentT.InvCDF(location, scale, dof, p), Is.EqualTo(x).Within(1e-6));
}
/// <summary>
/// Tau distribucija (razdioba)
/// </summary>
/// <param name="alfa">Nivo signifikantnosti (znacajnosti) -> 1 > alfa > 0</param>
/// <param name="f">Stupanj slobode</param>
/// <returns>double</returns>
/// <remarks>
/// <para/>Distribucija interno studentiziranih mjerenja
/// <para/>Popravci i njihove standardne su korelirani
/// <para/>Allan J. Pope 1976.
/// </remarks>
public static double Tau(double alfa, int f)
{
double t = StudentT.InvCDF(0, 1, f - 1, 1 - alfa / 2);
double tau = t * Math.Sqrt(f) / Math.Sqrt(f - 1 + Math.Pow(t, 2));
return(tau);
}
public double[] ConfidenceIntervalSimulationTime(double confidence)
{
confidence = 1 - ((1 - confidence) / 2);
double[] interval = new double[2];
double avg = _simulationTimeCumulative / _iterator;
double standardDeviation = Math.Sqrt((_simulationTimePower / _iterator) - Math.Pow(avg, 2));
double value = 0;
if (_iterator < 30)
{
value = StudentT.InvCDF(0.0, 1.0, _iterator, confidence);
}
else
{
value = Normal.InvCDF(0, 1, confidence);
}
interval[0] = avg - (value * standardDeviation / Math.Sqrt((_iterator - 1)));
interval[1] = avg + (value * standardDeviation / Math.Sqrt((_iterator - 1)));
return(interval);
}
public static double GetCriticalValue(double significanceLevel, double df)
{
return(StudentT.InvCDF(0.0, 1.0, df, 1.0 - significanceLevel / 2.0));
}
public static double TInv(double probability, int degreesFreedom)
{
return(-StudentT.InvCDF(0d, 1d, degreesFreedom, probability / 2));
}
/// <summary>
/// Inverzna Student (T) distribucija (razdioba)
/// </summary>
/// <param name="alfa">Nivo signifikantnosti (znacajnosti) -> 1 > alfa > 0</param>
/// <param name="f">Stupanj slobode</param>
/// <returns>double</returns>
public static double Student(double alfa, int f)
{
return(StudentT.InvCDF(0, 1, f, 1 - alfa));
}
public static double T_INV_2T(double p, double nu)
{
return(StudentT.InvCDF(location: 0.0, scale: 1.0, freedom: nu, p: 1.0 - p / 2.0));
}
private double ConfidenceInterval95(List <double> l, double standardError)
{
return(-StudentT.InvCDF(0, 1, l.Count - 1, 0.025) * standardError);
}