// Verify that distributions have the correct mean and variance.
// Note that sample mean and sample variance will not exactly match the expected mean and variance.
static void TestDistributions()
{
const int numSamples = 100000;
double mean, variance, stdev, shape, scale, degreesOfFreedom;
RunningStat rs = new RunningStat();
// Exponential distribution
rs.Clear();
mean = 2;
for (int i = 0; i < numSamples; ++i)
rs.Push(SimpleRNG.GetExponential(mean));
PrintResults("exponential", mean, 5, rs.Mean(), rs.Variance());
// Gamma distribution
rs.Clear();
shape = 10; scale = 2;
for (int i = 0; i < numSamples; ++i)
rs.Push(SimpleRNG.GetGamma(shape, scale));
PrintResults("gamma", shape*scale, shape*scale*scale, rs.Mean(), rs.Variance());
// Normal distribution
rs.Clear();
mean = 2; stdev = 5;
for (int i = 0; i < numSamples; ++i)
rs.Push(SimpleRNG.GetNormal(2, 5));
PrintResults("normal", mean, stdev*stdev, rs.Mean(), rs.Variance());
// Student t distribution
rs.Clear();
degreesOfFreedom = 6;
for (int i = 0; i < numSamples; ++i)
rs.Push(SimpleRNG.GetStudentT(6));
PrintResults("Student t", 0, degreesOfFreedom / (degreesOfFreedom - 2.0), rs.Mean(), rs.Variance());
// Weibull distribution
rs.Clear();
shape = 2; scale = 3;
mean = 3*Math.Sqrt(Math.PI)/2;
variance = 9*(1 - Math.PI/4);
for (int i = 0; i < numSamples; ++i)
rs.Push(SimpleRNG.GetWeibull(shape, scale));
PrintResults("Weibull", mean, variance, rs.Mean(), rs.Variance());
// Beta distribution
rs.Clear();
double a = 7, b = 2;
mean = a / (a + b);
variance = mean*(1 - mean) / (a + b + 1);
for (int i = 0; i < numSamples; ++i)
rs.Push(SimpleRNG.GetBeta(a, b));
PrintResults("Beta", mean, variance, rs.Mean(), rs.Variance());
// Verify that distributions have the correct mean and variance.
// Note that sample mean and sample variance will not exactly match the expected mean and variance.
static void TestDistributions()
{
const int numSamples = 100000;
double mean, variance, stdev, shape, scale, degreesOfFreedom;
RunningStat rs = new RunningStat();
// Exponential distribution
rs.Clear();
mean = 2;
for (int i = 0; i < numSamples; ++i)
{
rs.Push(SimpleRNG.GetExponential(mean));
}
PrintResults("exponential", mean, 5, rs.Mean(), rs.Variance());
// Gamma distribution
rs.Clear();
shape = 10; scale = 2;
for (int i = 0; i < numSamples; ++i)
{
rs.Push(SimpleRNG.GetGamma(shape, scale));
}
PrintResults("gamma", shape * scale, shape * scale * scale, rs.Mean(), rs.Variance());
// Normal distribution
rs.Clear();
mean = 2; stdev = 5;
for (int i = 0; i < numSamples; ++i)
{
rs.Push(SimpleRNG.GetNormal(2, 5));
}
PrintResults("normal", mean, stdev * stdev, rs.Mean(), rs.Variance());
// Student t distribution
rs.Clear();
degreesOfFreedom = 6;
for (int i = 0; i < numSamples; ++i)
{
rs.Push(SimpleRNG.GetStudentT(6));
}
PrintResults("Student t", 0, degreesOfFreedom / (degreesOfFreedom - 2.0), rs.Mean(), rs.Variance());
// Weibull distribution
rs.Clear();
shape = 2; scale = 3;
mean = 3 * Math.Sqrt(Math.PI) / 2;
variance = 9 * (1 - Math.PI / 4);
for (int i = 0; i < numSamples; ++i)
{
rs.Push(SimpleRNG.GetWeibull(shape, scale));
}
PrintResults("Weibull", mean, variance, rs.Mean(), rs.Variance());
// Beta distribution
rs.Clear();
double a = 7, b = 2;
mean = a / (a + b);
variance = mean * (1 - mean) / (a + b + 1);
for (int i = 0; i < numSamples; ++i)
{
rs.Push(SimpleRNG.GetBeta(a, b));
}
PrintResults("Beta", mean, variance, rs.Mean(), rs.Variance());
}
