public void FitTest2()
{
double[][] observations =
{
new double[] { 1, 2 },
new double[] { 1, 2 },
new double[] { 1, 2 },
new double[] { 1, 2 }
};
var target = new MultivariateNormalDistribution(2);
bool thrown = false;
try { target.Fit(observations); }
catch (NonPositiveDefiniteMatrixException) { thrown = true; }
Assert.IsTrue(thrown);
NormalOptions options = new NormalOptions() { Regularization = double.Epsilon };
// No exception thrown
target.Fit(observations, options);
}
public void EstimateTest()
{
double[] observations = { 2, 2, 2, 2, 2 };
NormalOptions options = new NormalOptions() { Regularization = 0.1 };
LognormalDistribution actual = LognormalDistribution.Estimate(observations, options);
Assert.AreEqual(System.Math.Log(2), actual.Location);
Assert.AreEqual(System.Math.Sqrt(0.1), actual.Shape);
}
public void FitTest3()
{
double[][] observations =
{
new double[] { 1, 2 },
new double[] { 2, 4 },
new double[] { 3, 6 },
new double[] { 4, 8 }
};
var target = new MultivariateNormalDistribution(2);
NormalOptions options = new NormalOptions()
{
Robust = true
};
target.Fit(observations, options);
double pdf = target.ProbabilityDensityFunction(4, 2);
double cdf = target.DistributionFunction(4, 2);
bool psd = target.Covariance.IsPositiveDefinite();
Assert.AreEqual(0.043239154739844896, pdf);
Assert.AreEqual(0.12263905840338646, cdf);
Assert.IsFalse(psd);
}
public void FitTest()
{
IDistribution target = CreateMultivariateContinuousDistribution(); // TODO: Initialize to an appropriate value
double[][] observations =
{
new double[] { 1, 1 },
new double[] { 1, 1 },
};
double[] weights = { 0.50, 0.50 };
bool thrown;
IFittingOptions options = new NormalOptions()
{
Regularization = 0.1
};
thrown = false;
try
{
target.Fit(observations, weights);
}
catch
{
thrown = true;
}
Assert.AreEqual(true, thrown);
thrown = false;
try
{
target.Fit(observations, weights, options);
}
catch
{
thrown = true;
}
Assert.AreEqual(false, thrown);
}
public void FitTest5()
{
double[][] observations =
{
new double[] { 1, 2 },
new double[] { 1, 2 },
new double[] { 0, 1 },
new double[] { 5, 7 }
};
double[] weights = { 1, 1, 0, 0 };
var target = new MultivariateNormalDistribution(2);
bool thrown = false;
try { target.Fit(observations, weights); }
catch (NonPositiveDefiniteMatrixException) { thrown = true; }
Assert.IsTrue(thrown);
NormalOptions options = new NormalOptions() { Robust = true };
// No exception thrown
target.Fit(observations, weights, options);
checkDegenerate(target);
}
/// <summary>
/// Estimates a new Normal distribution from a given set of observations.
/// </summary>
///
public static NormalDistribution Estimate(double[] observations, double[] weights, NormalOptions options)
{
NormalDistribution n = new NormalDistribution();
n.Fit(observations, weights, options);
return n;
}
/// <summary>
/// Estimates a new Normal distribution from a given set of observations.
/// </summary>
///
public static NormalDistribution Estimate(double[] observations, NormalOptions options)
{
return Estimate(observations, null, options);
}
/// <summary>
/// Estimates a new Normal distribution from a given set of observations.
/// </summary>
///
public static MultivariateNormalDistribution Estimate(double[][] observations, double[] weights, NormalOptions options)
{
MultivariateNormalDistribution n = new MultivariateNormalDistribution(observations[0].Length);
n.Fit(observations, weights, options);
return n;
}
/// <summary>
/// Creates a Baum-Welch with default configurations for
/// hidden Markov models with normal mixture densities.
/// </summary>
///
public static BaumWelchLearning<MultivariateMixture<MultivariateNormalDistribution>> FromMixtureModel(
HiddenMarkovModel<MultivariateMixture<MultivariateNormalDistribution>> model, NormalOptions options)
{
MixtureOptions mixOptions = new MixtureOptions()
{
Iterations = 1,
InnerOptions = options
};
return new BaumWelchLearning<MultivariateMixture<MultivariateNormalDistribution>>(model)
{
FittingOptions = mixOptions
};
}
