public void TestMethod1()
{
Distribution n0 = new TransformedDistribution(new NormalDistribution(), -2.0, 3.0);
Distribution n1 = new NormalDistribution(-2.0, 3.0);
Assert.IsTrue(TestUtilities.IsNearlyEqual(n0.Mean, n1.Mean));
Assert.IsTrue(TestUtilities.IsNearlyEqual(n0.Variance, n1.Variance));
Assert.IsTrue(TestUtilities.IsNearlyEqual(n0.StandardDeviation, n1.StandardDeviation));
Assert.IsTrue(TestUtilities.IsNearlyEqual(n0.Skewness, n1.Skewness));
for (int k = 0; k < 8; k++) {
Assert.IsTrue(TestUtilities.IsNearlyEqual(n0.Moment(k), n1.Moment(k)));
Assert.IsTrue(TestUtilities.IsNearlyEqual(n0.MomentAboutMean(k), n1.MomentAboutMean(k)));
}
foreach (double x in TestUtilities.GenerateUniformRealValues(-8.0, 8.0, 8)) {
Assert.IsTrue(TestUtilities.IsNearlyEqual(n0.ProbabilityDensity(x), n1.ProbabilityDensity(x)));
Assert.IsTrue(TestUtilities.IsNearlyEqual(n0.LeftProbability(x), n1.LeftProbability(x)));
Assert.IsTrue(TestUtilities.IsNearlyEqual(n0.RightProbability(x), n1.RightProbability(x)));
}
foreach (double P in TestUtilities.GenerateRealValues(1.0E-4, 1.0, 4)) {
Assert.IsTrue(TestUtilities.IsNearlyEqual(n0.InverseLeftProbability(P), n1.InverseLeftProbability(P)));
Assert.IsTrue(TestUtilities.IsNearlyEqual(n0.InverseRightProbability(P), n1.InverseRightProbability(P)));
}
}
public void AnovaDistribution()
{
Distribution sDistribution = new NormalDistribution();
Random rng = new Random(1);
Sample fSample = new Sample();
// do 100 ANOVAs
for (int t = 0; t < 100; t++) {
// each ANOVA has 4 groups
List<Sample> groups = new List<Sample>();
for (int g = 0; g < 4; g++) {
// each group has 3 data points
Sample group = new Sample();
for (int i = 0; i < 3; i++) {
group.Add(sDistribution.GetRandomValue(rng));
}
groups.Add(group);
}
OneWayAnovaResult result = Sample.OneWayAnovaTest(groups);
fSample.Add(result.Factor.Result.Statistic);
}
// compare the distribution of F statistics to the expected distribution
Distribution fDistribution = new FisherDistribution(3, 8);
Console.WriteLine("m={0} s={1}", fSample.PopulationMean, fSample.PopulationStandardDeviation);
TestResult kResult = fSample.KolmogorovSmirnovTest(fDistribution);
Console.WriteLine(kResult.LeftProbability);
Assert.IsTrue(kResult.LeftProbability < 0.95);
}
public MultivariateSample CreateMultivariateNormalSample(ColumnVector M, SymmetricMatrix C, int n)
{
int d = M.Dimension;
MultivariateSample S = new MultivariateSample(d);
SquareMatrix A = C.CholeskyDecomposition().SquareRootMatrix();
Random rng = new Random(1);
Distribution normal = new NormalDistribution();
for (int i = 0; i < n; i++) {
// create a vector of normal deviates
ColumnVector V = new ColumnVector(d);
for (int j = 0; j < d; j++) {
double y = rng.NextDouble();
double z = normal.InverseLeftProbability(y);
V[j] = z;
}
// form the multivariate distributed vector
ColumnVector X = M + A * V;
// add it to the sample
S.Add(X);
}
return (S);
}
/*
* x: mide el tiempo
* y: mide la cantidad que ingresa en el anden
* m: es el máximo de la campana de gauss
* r1, r2: son números aleatorios
*/
public static int Normal(int minValue, int maxValue)
{
double r1, r2, m, x, y, mu, sigma, fx;
if ((minValue >= maxValue) || (minValue < 0) || (maxValue < 0))
{
throw new System.ArgumentException("Valores incorrectos");
}
mu = calcularMedia(minValue, maxValue);
sigma = calcularDesvio(minValue, maxValue);
NormalDistribution f = new NormalDistribution(mu, sigma);
m = 1 / (sigma * Math.Sqrt(2 * Math.PI));
do
{
r1 = Rand();
r2 = Rand();
x = minValue + (maxValue - minValue) * r1;
y = m * r2;
fx = f.ProbabilityDensity((x - mu) / sigma);
} while (y <= fx);
return Convert.ToInt32(x);
}
public static void Mutate(Chromosome chromosome, Random random)
{
NormalDistribution normal;
UniformDistribution uniform = new UniformDistribution(Interval.FromEndpoints(0, 1));
for (int i = 0; i < chromosome.Values.Count; i++)
{
if (uniform.GetRandomValue(random) <= MUTATION_PROBABILITY)
{
if (uniform.GetRandomValue(random) <= SELECTION_PROBABILITY)
normal = new NormalDistribution(MEAN, SIGMA_1);
else
normal = new NormalDistribution(MEAN, SIGMA_2);
chromosome.Values[i] += normal.GetRandomValue(random);
}
}
}
public void TwoSampleKS2()
{
int n = 2 * 3 * 3; int m = 2 * 2 * 3;
Random rng = new Random(0);
NormalDistribution d = new NormalDistribution();
Histogram h = new Histogram((int) AdvancedIntegerMath.LCM(n, m) + 1);
//int[] h = new int[AdvancedIntegerMath.LCM(n, m) + 1];
int count = 1000;
for (int i = 0; i < count; i++) {
Sample A = new Sample();
for (int j = 0; j < n; j++) A.Add(d.GetRandomValue(rng));
Sample B = new Sample();
for (int j = 0; j < m; j++) B.Add(d.GetRandomValue(rng));
TestResult r = Sample.KolmogorovSmirnovTest(A, B);
int k = (int) Math.Round(r.Statistic * AdvancedIntegerMath.LCM(n, m));
//Console.WriteLine("{0} {1}", r.Statistic, k);
h[k].Increment();
//h[k] = h[k] + 1;
}
KolmogorovTwoSampleExactDistribution ks = new KolmogorovTwoSampleExactDistribution(n, m);
double chi2 = 0.0; int dof = 0;
for (int i = 0; i < h.Count; i++) {
double ne = ks.ProbabilityMass(i) * count;
Console.WriteLine("{0} {1} {2}", i, h[i].Counts, ne);
if (ne > 4) {
chi2 += MoreMath.Sqr(h[i].Counts - ne) / ne;
dof++;
}
}
Console.WriteLine("chi^2={0} dof={1}", chi2, dof);
TestResult r2 = h.ChiSquaredTest(ks);
ChiSquaredDistribution rd = (ChiSquaredDistribution) r2.Distribution;
Console.WriteLine("chi^2={0} dof={1} P={2}", r2.Statistic, rd.DegreesOfFreedom, r2.RightProbability);
}
public void SampleMaximumLikelihoodFit()
{
// normal distriubtion
Console.WriteLine("normal");
double mu = -1.0;
double sigma = 2.0;
Distribution nd = new NormalDistribution(mu, sigma);
Sample ns = CreateSample(nd, 500);
//FitResult nr = ns.MaximumLikelihoodFit(new NormalDistribution(mu + 1.0, sigma + 1.0));
FitResult nr = ns.MaximumLikelihoodFit((IList<double> p) => new NormalDistribution(p[0], p[1]), new double[] { mu + 1.0, sigma + 1.0 });
Console.WriteLine(nr.Parameter(0));
Console.WriteLine(nr.Parameter(1));
Assert.IsTrue(nr.Dimension == 2);
Assert.IsTrue(nr.Parameter(0).ConfidenceInterval(0.95).ClosedContains(mu));
Assert.IsTrue(nr.Parameter(1).ConfidenceInterval(0.95).ClosedContains(sigma));
FitResult nr2 = NormalDistribution.FitToSample(ns);
Console.WriteLine(nr.Covariance(0,1));
// test analytic expression
Assert.IsTrue(TestUtilities.IsNearlyEqual(nr.Parameter(0).Value, ns.Mean, Math.Sqrt(TestUtilities.TargetPrecision)));
// we don't expect to be able to test sigma against analytic expression because ML value has known bias for finite sample size
// exponential distribution
Console.WriteLine("exponential");
double em = 3.0;
Distribution ed = new ExponentialDistribution(em);
Sample es = CreateSample(ed, 100);
//FitResult er = es.MaximumLikelihoodFit(new ExponentialDistribution(em + 1.0));
FitResult er = es.MaximumLikelihoodFit((IList<double> p) => new ExponentialDistribution(p[0]), new double[] { em + 1.0 });
Console.WriteLine(er.Parameter(0));
Assert.IsTrue(er.Dimension == 1);
Assert.IsTrue(er.Parameter(0).ConfidenceInterval(0.95).ClosedContains(em));
// test against analytic expression
Assert.IsTrue(TestUtilities.IsNearlyEqual(er.Parameter(0).Value, es.Mean, Math.Sqrt(TestUtilities.TargetPrecision)));
Assert.IsTrue(TestUtilities.IsNearlyEqual(er.Parameter(0).Uncertainty, es.Mean / Math.Sqrt(es.Count), Math.Sqrt(Math.Sqrt(TestUtilities.TargetPrecision))));
// lognormal distribution
Console.WriteLine("lognormal");
double l1 = -4.0;
double l2 = 5.0;
Distribution ld = new LognormalDistribution(l1, l2);
Sample ls = CreateSample(ld, 100);
//FitResult lr = ls.MaximumLikelihoodFit(new LognormalDistribution(l1 + 1.0, l2 + 1.0));
FitResult lr = ls.MaximumLikelihoodFit((IList<double> p) => new LognormalDistribution(p[0], p[1]), new double[] { l1 + 1.0, l2 + 1.0 });
Console.WriteLine(lr.Parameter(0));
Console.WriteLine(lr.Parameter(1));
Console.WriteLine(lr.Covariance(0, 1));
Assert.IsTrue(lr.Dimension == 2);
Assert.IsTrue(lr.Parameter(0).ConfidenceInterval(0.99).ClosedContains(l1));
Assert.IsTrue(lr.Parameter(1).ConfidenceInterval(0.99).ClosedContains(l2));
// weibull distribution
Console.WriteLine("weibull");
double w_scale = 4.0;
double w_shape = 2.0;
WeibullDistribution w_d = new WeibullDistribution(w_scale, w_shape);
Sample w_s = CreateSample(w_d, 20);
//FitResult w_r = w_s.MaximumLikelihoodFit(new WeibullDistribution(1.0, 0.5));
FitResult w_r = w_s.MaximumLikelihoodFit((IList<double> p) => new WeibullDistribution(p[0], p[1]), new double[] { 2.0, 2.0 });
Console.WriteLine(w_r.Parameter(0));
Console.WriteLine(w_r.Parameter(1));
Console.WriteLine(w_r.Covariance(0, 1));
Assert.IsTrue(w_r.Parameter(0).ConfidenceInterval(0.95).ClosedContains(w_d.ScaleParameter));
Assert.IsTrue(w_r.Parameter(1).ConfidenceInterval(0.95).ClosedContains(w_d.ShapeParameter));
// logistic distribution
Console.WriteLine("logistic");
double logistic_m = -3.0;
double logistic_s = 2.0;
Distribution logistic_distribution = new LogisticDistribution(logistic_m, logistic_s);
Sample logistic_sample = CreateSample(logistic_distribution, 100);
//FitResult logistic_result = logistic_sample.MaximumLikelihoodFit(new LogisticDistribution());
FitResult logistic_result = logistic_sample.MaximumLikelihoodFit((IList<double> p) => new LogisticDistribution(p[0], p[1]), new double[] { 2.0, 3.0 });
Console.WriteLine(logistic_result.Parameter(0));
Console.WriteLine(logistic_result.Parameter(1));
Assert.IsTrue(logistic_result.Dimension == 2);
Assert.IsTrue(logistic_result.Parameter(0).ConfidenceInterval(0.95).ClosedContains(logistic_m));
Assert.IsTrue(logistic_result.Parameter(1).ConfidenceInterval(0.95).ClosedContains(logistic_s));
// beta distribution
// not yet!
/*
double beta_alpha = 0.5;
double beta_beta = 2.0;
Distribution beta_distribution = new BetaDistribution(beta_alpha, beta_beta);
Sample beta_sample = CreateSample(beta_distribution, 100);
FitResult beta_result = beta_sample.MaximumLikelihoodFit(new BetaDistribution(1.0, 1.0));
Console.WriteLine("Beta:");
Console.WriteLine(beta_result.Parameter(0));
Console.WriteLine(beta_result.Parameter(1));
Assert.IsTrue(beta_result.Dimension == 2);
Assert.IsTrue(beta_result.Parameter(0).ConfidenceInterval(0.95).ClosedContains(beta_alpha));
Assert.IsTrue(beta_result.Parameter(1).ConfidenceInterval(0.95).ClosedContains(beta_beta));
*/
}
public void NormalFitUncertainties()
{
NormalDistribution N = new NormalDistribution(-1.0, 2.0);
// Create a bivariate sample to hold our fitted best mu and sigma values
// so we can determine their covariance as well as their means and variances
BivariateSample fits = new BivariateSample();
double cmm = 0.0, css = 0.0, cms = 0.0;
// A bunch of times, create a normal sample
for (int i = 0; i < 64; i++) {
// we will use small samples so the variation in mu and sigma will be more substantial
Sample s = TestUtilities.CreateSample(N, 16, i);
// fit each sample to a normal distribution
FitResult fit = NormalDistribution.FitToSample(s);
// and record the mu and sigma values from the fit into our bivariate sample
fits.Add(fit.Parameter(0).Value, fit.Parameter(1).Value);
// also record the claimed covariances among these parameters
cmm += fit.Covariance(0, 0); css += fit.Covariance(1, 1); cms += fit.Covariance(0, 1);
}
cmm /= fits.Count; css /= fits.Count; cms /= fits.Count;
// the mean fit values should agree with the population distribution
Console.WriteLine("{0} {1}", fits.X.PopulationMean, N.Mean);
Assert.IsTrue(fits.X.PopulationMean.ConfidenceInterval(0.95).ClosedContains(N.Mean));
Console.WriteLine("{0} {1}", fits.Y.PopulationMean, N.StandardDeviation);
Assert.IsTrue(fits.Y.PopulationMean.ConfidenceInterval(0.95).ClosedContains(N.StandardDeviation));
// but also the covariances of those fit values should agree with the claimed covariances
Console.WriteLine("{0} {1}", fits.X.PopulationVariance, cmm);
Assert.IsTrue(fits.X.PopulationVariance.ConfidenceInterval(0.95).ClosedContains(cmm));
Console.WriteLine("{0} {1}", fits.Y.PopulationVariance, css);
Assert.IsTrue(fits.Y.PopulationVariance.ConfidenceInterval(0.95).ClosedContains(css));
Console.WriteLine("{0} {1}", fits.PopulationCovariance, cms);
Assert.IsTrue(fits.PopulationCovariance.ConfidenceInterval(0.95).ClosedContains(cms));
/*
Random rng = new Random(2718281);
BivariateSample P = new BivariateSample();
double cmm = 0.0;
double css = 0.0;
double cms = 0.0;
for (int i = 0; i < 64; i++) {
Sample s = new Sample();
for (int j = 0; j < 16; j++) {
s.Add(N.GetRandomValue(rng));
}
FitResult r = NormalDistribution.FitToSample(s);
P.Add(r.Parameter(0).Value, r.Parameter(1).Value);
cmm += r.Covariance(0, 0);
css += r.Covariance(1, 1);
cms += r.Covariance(0, 1);
}
cmm /= P.Count;
css /= P.Count;
cms /= P.Count;
Console.WriteLine("{0} {1}", P.X.PopulationMean, P.Y.PopulationMean);
Assert.IsTrue(P.X.PopulationMean.ConfidenceInterval(0.95).ClosedContains(N.Mean));
Assert.IsTrue(P.Y.PopulationMean.ConfidenceInterval(0.95).ClosedContains(N.StandardDeviation));
Assert.IsTrue(P.X.PopulationVariance.ConfidenceInterval(0.95).ClosedContains(cmm));
Assert.IsTrue(P.Y.PopulationVariance.ConfidenceInterval(0.95).ClosedContains(css));
Assert.IsTrue(P.PopulationCovariance.ConfidenceInterval(0.95).ClosedContains(cms));
*/
}
public void PearsonRDistribution()
{
Random rng = new Random(1);
// pick some underlying distributions for the sample variables, which must be normal but can have any parameters
NormalDistribution xDistribution = new NormalDistribution(1, 2);
NormalDistribution yDistribution = new NormalDistribution(3, 4);
// try this for several sample sizes, all low so that we see the difference from the normal distribution
// n = 3 maxima at ends; n = 4 uniform; n = 5 semi-circular "mound"; n = 6 parabolic "mound"
foreach (int n in new int[] { 3, 4, 5, 6, 8 }) {
Console.WriteLine("n={0}", n);
// find r values
Sample rSample = new Sample();
for (int i = 0; i < 100; i++) {
// to get each r value, construct a bivariate sample of the given size with no cross-correlation
BivariateSample xySample = new BivariateSample();
for (int j = 0; j < n; j++) {
xySample.Add(xDistribution.GetRandomValue(rng), yDistribution.GetRandomValue(rng));
}
double r = xySample.PearsonRTest().Statistic;
rSample.Add(r);
}
// check whether r is distributed as expected
TestResult result = rSample.KolmogorovSmirnovTest(new PearsonRDistribution(n));
Console.WriteLine("P={0}", result.LeftProbability);
Assert.IsTrue(result.LeftProbability < 0.95);
}
}
public void BivariateLinearRegressionGoodnessOfFitDistribution()
{
// create uncorrelated x and y values
// the distribution of F-test statistics returned by linear fits should follow the expected F-distribution
Random rng = new Random(987654321);
NormalDistribution xd = new NormalDistribution(1.0, 2.0);
NormalDistribution yd = new NormalDistribution(-3.0, 4.0);
Sample fs = new Sample();
for (int i = 0; i < 127; i++) {
BivariateSample xys = new BivariateSample();
for (int j = 0; j < 7; j++) {
xys.Add(xd.GetRandomValue(rng), yd.GetRandomValue(rng));
}
double f = xys.LinearRegression().GoodnessOfFit.Statistic;
fs.Add(f);
}
Distribution fd = new FisherDistribution(1, 5);
Console.WriteLine("{0} v. {1}", fs.PopulationMean, fd.Mean);
TestResult t = fs.KolmogorovSmirnovTest(fd);
Console.WriteLine(t.LeftProbability);
Assert.IsTrue(t.LeftProbability < 0.95);
}
public void StudentTest2()
{
// make sure Student t is consistent with its definition
// we are going to take a sample that we expect to be t-distributed
Sample tSample = new Sample();
// begin with an underlying normal distribution
Distribution xDistribution = new NormalDistribution();
// compute a bunch of t satistics from the distribution
for (int i = 0; i < 100000; i++) {
// take a small sample from the underlying distribution
// (as the sample gets large, the t distribution becomes normal)
Random rng = new Random(314159+i);
double p = xDistribution.InverseLeftProbability(rng.NextDouble());
double q = 0.0;
for (int j = 0; j < 5; j++) {
double x = xDistribution.InverseLeftProbability(rng.NextDouble());
q += x * x;
}
q = q / 5;
double t = p / Math.Sqrt(q);
tSample.Add(t);
}
Distribution tDistribution = new StudentDistribution(5);
TestResult result = tSample.KolmogorovSmirnovTest(tDistribution);
Console.WriteLine(result.LeftProbability);
}
public void TestNormalOrderStatistic()
{
int n = 100;
//int r = 3 * n / 4;
int r = 52;
Distribution d = new NormalDistribution();
double C = Math.Exp(AdvancedIntegerMath.LogFactorial(n) - AdvancedIntegerMath.LogFactorial(r - 1) - AdvancedIntegerMath.LogFactorial(n - r));
double m = GaussHermiteIntegrate(x => C * MoreMath.Pow(d.LeftProbability(x), r - 1) * MoreMath.Pow(d.RightProbability(x), n - r) * x);
//double m = GaussHermiteIntegrate(x => 1.0);
double m2 = FunctionMath.Integrate(
//x => 1.0 * Math.Exp(-x * x / 2.0) / Math.Sqrt(2.0 * Math.PI),
x => C * MoreMath.Pow(d.LeftProbability(x), r - 1) * MoreMath.Pow(d.RightProbability(x), n - r) * x * Math.Exp(-x * x / 2.0) / Math.Sqrt(2.0 * Math.PI),
Interval.FromEndpoints(Double.NegativeInfinity, Double.PositiveInfinity)
);
Console.WriteLine(m);
Console.WriteLine(m2);
Console.WriteLine(NormalMeanOrderStatisticExpansion(r, n));
Console.WriteLine(NormalMeanOrderStatisticExpansion2(r, n));
//Console.WriteLine(1.5 / Math.Sqrt(Math.PI));
}
public void TestBeta()
{
double a = 200.0; double b = 200.0; double P = 1.0E-5;
double x1 = ApproximateInverseBetaSeries(a, b, P);
if ((0.0 < x1) && (x1 < 1.0)) {
Console.WriteLine("x1 {0} {1}", x1, AdvancedMath.LeftRegularizedBeta(a, b, x1));
}
double x2 = 1.0 - ApproximateInverseBetaSeries(b, a, 1.0 - P);
if ((0.0 < x2) && (x2 < 1.0)) {
Console.WriteLine("x2 {0} {1}", x2, AdvancedMath.LeftRegularizedBeta(a, b, x2));
}
//x1 = RefineInverseBeta(a, b, P, x1);
//Console.WriteLine("{0} {1}", x1, AdvancedMath.LeftRegularizedBeta(a, b, x1));
NormalDistribution N = new NormalDistribution();
double m = a / (a + b); double s = Math.Sqrt(a * b / (a + b + 1.0)) / (a + b);
double x3 = m + s * N.InverseLeftProbability(P);
if ((0.0 < x3) && (x3 < 1.0)) {
Console.WriteLine("x3 {0} {1}", x3, AdvancedMath.LeftRegularizedBeta(a, b, x3));
}
//Console.WriteLine(AdvancedMath.Beta(a, b, 0.35) / AdvancedMath.Beta(a, b));
//Console.WriteLine(AdvancedMath.Beta(a, b, 0.40) / AdvancedMath.Beta(a, b));
//Console.WriteLine(AdvancedMath.Beta(a, b, 0.45) / AdvancedMath.Beta(a, b));
}
public void KuiperNullDistributionTest()
{
// The distribution is irrelevent; pick one at random
Distribution sampleDistribution = new NormalDistribution();
// Loop over various sample sizes
foreach (int n in TestUtilities.GenerateIntegerValues(2, 128, 16)) {
// Create a sample to hold the KS statistics
Sample testStatistics = new Sample();
// and a variable to hold the claimed null distribution, which should be the same for each test
Distribution nullDistribution = null;
// Create a bunch of samples, each with n+1 data points
// We pick n+1 instead of n just to have different sample size values than in the KS test case
for (int i = 0; i < 256; i++) {
// Just use n+i as a seed in order to get different points each time
Sample sample = TestUtilities.CreateSample(sampleDistribution, n + 1, 512 * n + i + 2);
// Do a Kuiper test of the sample against the distribution each time
TestResult r1 = sample.KuiperTest(sampleDistribution);
// Record the test statistic value and the claimed null distribution
testStatistics.Add(r1.Statistic);
nullDistribution = r1.Distribution;
}
// Do a KS test of our sample of Kuiper statistics against the claimed null distribution
// We could use a Kuiper test here instead, which would be way cool and meta, but we picked KS instead for variety
TestResult r2 = testStatistics.KolmogorovSmirnovTest(nullDistribution);
Console.WriteLine("{0} {1} {2}", n, r2.Statistic, r2.LeftProbability);
Assert.IsTrue(r2.RightProbability > 0.01);
// Test moment matches, too
Console.WriteLine(" {0} {1}", testStatistics.PopulationMean, nullDistribution.Mean);
Console.WriteLine(" {0} {1}", testStatistics.PopulationVariance, nullDistribution.Variance);
Assert.IsTrue(testStatistics.PopulationMean.ConfidenceInterval(0.99).ClosedContains(nullDistribution.Mean));
Assert.IsTrue(testStatistics.PopulationVariance.ConfidenceInterval(0.99).ClosedContains(nullDistribution.Variance));
}
}
public void MultivariateMoments()
{
// create a random sample
MultivariateSample M = new MultivariateSample(3);
Distribution d0 = new NormalDistribution();
Distribution d1 = new ExponentialDistribution();
Distribution d2 = new UniformDistribution();
Random rng = new Random(1);
int n = 10;
for (int i = 0; i < n; i++) {
M.Add(d0.GetRandomValue(rng), d1.GetRandomValue(rng), d2.GetRandomValue(rng));
}
// test that moments agree
for (int i = 0; i < 3; i++) {
int[] p = new int[3];
p[i] = 1;
Assert.IsTrue(TestUtilities.IsNearlyEqual(M.Column(i).Mean, M.Moment(p)));
p[i] = 2;
Assert.IsTrue(TestUtilities.IsNearlyEqual(M.Column(i).Variance, M.MomentAboutMean(p)));
for (int j = 0; j < i; j++) {
int[] q = new int[3];
q[i] = 1;
q[j] = 1;
Assert.IsTrue(TestUtilities.IsNearlyEqual(M.TwoColumns(i, j).Covariance, M.MomentAboutMean(q)));
}
}
}
public void MultivariateLinearRegressionTest()
{
// define model y = a + b0 * x0 + b1 * x1 + noise
double a = 1.0;
double b0 = -2.0;
double b1 = 3.0;
Distribution noise = new NormalDistribution(0.0, 10.0);
// draw a sample from the model
Random rng = new Random(1);
MultivariateSample sample = new MultivariateSample(3);
for (int i = 0; i < 100; i++) {
double x0 = -10.0 + 20.0 * rng.NextDouble();
double x1 = -10.0 + 20.0 * rng.NextDouble();
double eps = noise.InverseLeftProbability(rng.NextDouble());
double y = a + b0 * x0 + b1 * x1 + eps;
sample.Add(x0, x1, y);
}
// do a linear regression fit on the model
FitResult result = sample.LinearRegression(2);
// the result should have the appropriate dimension
Assert.IsTrue(result.Dimension == 3);
// the result should be significant
Console.WriteLine("{0} {1}", result.GoodnessOfFit.Statistic, result.GoodnessOfFit.LeftProbability);
Assert.IsTrue(result.GoodnessOfFit.LeftProbability > 0.95);
// the parameters should match the model
Console.WriteLine(result.Parameter(0));
Assert.IsTrue(result.Parameter(0).ConfidenceInterval(0.90).ClosedContains(b0));
Console.WriteLine(result.Parameter(1));
Assert.IsTrue(result.Parameter(1).ConfidenceInterval(0.90).ClosedContains(b1));
Console.WriteLine(result.Parameter(2));
Assert.IsTrue(result.Parameter(2).ConfidenceInterval(0.90).ClosedContains(a));
}
public void MultivariateLinearRegressionNullDistribution()
{
int d = 4;
Random rng = new Random(1);
NormalDistribution n = new NormalDistribution();
Sample fs = new Sample();
for (int i = 0; i < 64; i++) {
MultivariateSample ms = new MultivariateSample(d);
for (int j = 0; j < 8; j++) {
double[] x = new double[d];
for (int k = 0; k < d; k++) {
x[k] = n.GetRandomValue(rng);
}
ms.Add(x);
}
FitResult r = ms.LinearRegression(0);
fs.Add(r.GoodnessOfFit.Statistic);
}
// conduct a KS test to check that F follows the expected distribution
TestResult ks = fs.KolmogorovSmirnovTest(new FisherDistribution(3, 4));
Assert.IsTrue(ks.LeftProbability < 0.95);
}
/// <summary>
/// Computes the normal distribution that best fits the given sample.
/// </summary>
/// <param name="sample">The sample to fit.</param>
/// <returns>The best fit parameters.</returns>
/// <remarks>
/// <para>The returned fit parameters are the μ (<see cref="Mean"/>) and σ (<see cref="StandardDeviation"/>) parameters, in that order.
/// These are the same parameters, in the same order, that are required by the <see cref="NormalDistribution(double,double)"/> constructor to
/// specify a new normal distribution.</para>
/// </remarks>
/// <exception cref="ArgumentNullException"><paramref name="sample"/> is null.</exception>
/// <exception cref="InsufficientDataException"><paramref name="sample"/> contains fewer than three values.</exception>
public static FitResult FitToSample(Sample sample)
{
if (sample == null) throw new ArgumentNullException("sample");
if (sample.Count < 3) throw new InsufficientDataException();
// maximum likelyhood estimates are guaranteed to be asymptotically unbiased, but not necessarily unbiased
// this hits home for the maximum likelyhood estimate of the variance of a normal distribution, which fails
// to include the N/(N-1) correction factor. since we know the bias, there is no reason for us not to correct
// it, and we do so here
UncertainValue mu = sample.PopulationMean;
UncertainValue sigma = sample.PopulationStandardDeviation;
Distribution distribution = new NormalDistribution(mu.Value, sigma.Value);
TestResult test = sample.KolmogorovSmirnovTest(distribution);
// the best-fit sigma and mu are known to be uncorrelated
// you can prove this by writing down the log likelyhood function and
// computing its mixed second derivative, which you will see vanishes
// at the minimum
return (new FitResult(mu.Value, mu.Uncertainty, sigma.Value, sigma.Uncertainty, 0.0, test));
}
public void TestNormal()
{
NormalDistribution n = new NormalDistribution();
Console.WriteLine(n.InverseLeftProbability(1.0E-10));
Console.WriteLine(n.InverseRightProbability(1.0E-10));
Console.WriteLine(n.InverseLeftProbability(1.0E-300));
Console.WriteLine(n.InverseRightProbability(1.0E-300));
Console.WriteLine(n.InverseLeftProbability(1.0));
//Console.WriteLine(n.InverseLeftProbability(0.26));
}
public void TTestDistribution()
{
// start with a normally distributed population
Distribution xDistribution = new NormalDistribution(2.0, 3.0);
Random rng = new Random(1);
// draw 100 samples from it and compute the t statistic for each
Sample tSample = new Sample();
for (int i = 0; i < 100; i++) {
// each sample has 9 values
Sample xSample = new Sample();
for (int j = 0; j < 9; j++) {
xSample.Add(xDistribution.GetRandomValue(rng));
}
//Sample xSample = CreateSample(xDistribution, 10, i);
TestResult tResult = xSample.StudentTTest(2.0);
double t = tResult.Statistic;
Console.WriteLine("t = {0}", t);
tSample.Add(t);
}
// sanity check our sample of t's
Assert.IsTrue(tSample.Count == 100);
// check that the t statistics are distributed as expected
Distribution tDistribution = new StudentDistribution(9);
// check on the mean
Console.WriteLine("m = {0} vs. {1}", tSample.PopulationMean, tDistribution.Mean);
Assert.IsTrue(tSample.PopulationMean.ConfidenceInterval(0.95).ClosedContains(tDistribution.Mean), String.Format("{0} vs. {1}", tSample.PopulationMean, tDistribution.Mean));
// check on the standard deviation
Console.WriteLine("s = {0} vs. {1}", tSample.PopulationStandardDeviation, tDistribution.StandardDeviation);
Assert.IsTrue(tSample.PopulationStandardDeviation.ConfidenceInterval(0.95).ClosedContains(tDistribution.StandardDeviation));
// do a KS test
TestResult ksResult = tSample.KolmogorovSmirnovTest(tDistribution);
Assert.IsTrue(ksResult.LeftProbability < 0.95);
Console.WriteLine("D = {0}", ksResult.Statistic);
// check that we can distinguish the t distribution from a normal distribution?
}
public void StudentTest()
{
// make sure Student t is consistent with its definition
// we are going to take a sample that we expect to be t-distributed
Sample tSample = new Sample();
// begin with an underlying normal distribution
Distribution xDistribution = new NormalDistribution(1.0, 2.0);
// compute a bunch of t satistics from the distribution
for (int i = 0; i < 200000; i++) {
// take a small sample from the underlying distribution
// (as the sample gets large, the t distribution becomes normal)
Random rng = new Random(i);
Sample xSample = new Sample();
for (int j = 0; j < 5; j++) {
double x = xDistribution.InverseLeftProbability(rng.NextDouble());
xSample.Add(x);
}
// compute t for the sample
double t = (xSample.Mean - xDistribution.Mean) / (xSample.PopulationStandardDeviation.Value / Math.Sqrt(xSample.Count));
tSample.Add(t);
//Console.WriteLine(t);
}
// t's should be t-distrubuted; use a KS test to check this
Distribution tDistribution = new StudentDistribution(4);
TestResult result = tSample.KolmogorovSmirnovTest(tDistribution);
Console.WriteLine(result.LeftProbability);
//Assert.IsTrue(result.LeftProbability < 0.95);
// t's should be demonstrably not normally distributed
Console.WriteLine(tSample.KolmogorovSmirnovTest(new NormalDistribution()).LeftProbability);
//Assert.IsTrue(tSample.KolmogorovSmirnovTest(new NormalDistribution()).LeftProbability > 0.95);
}
public void ZTestDistribution()
{
Random rng = new Random(1);
// define the sampling population (which must be normal for a z-test)
Distribution population = new NormalDistribution(2.0, 3.0);
// collect 100 samples
Sample zSample = new Sample();
for (int i = 0; i < 100; i++) {
// each z-statistic is formed by making a 4-count sample from a normal distribution
Sample sample = new Sample();
for (int j = 0; j < 4; j++) {
sample.Add(population.GetRandomValue(rng));
}
// for each sample, do a z-test against the population
TestResult zResult = sample.ZTest(population.Mean, population.StandardDeviation);
zSample.Add(zResult.Statistic);
}
// the z's should be distrubuted normally
TestResult result = zSample.KolmogorovSmirnovTest(new NormalDistribution());
Console.WriteLine("{0} {1}", result.Statistic, result.LeftProbability);
Assert.IsTrue((result.LeftProbability > 0.05) && (result.LeftProbability < 0.95));
}
public void MomentMapTest()
{
Distribution d = new NormalDistribution();
for (int n = 1; n < 11; n++) {
double[] K = new double[n+1];
K[0] = 1.0;
if (K.Length > 1) K[1] = 0.0;
if (K.Length > 2) K[2] = 1.0;
//for (int m = 1; m < K.Length; m++) {
// K[m] = AdvancedIntegerMath.Factorial(m - 1);
//}
double M = MomentMath.RawMomentFromCumulants(K);
Console.WriteLine("{0} {1}", d.Moment(n), M);
}
}
public void MaximumLikelihoodFitToNormal()
{
// create a normal sample
double mu = -1.0;
double sigma = 2.0;
Distribution d = new NormalDistribution(mu, sigma);
Sample s = CreateSample(d, 1024);
// do an explicit maximum likelyhood fit to a normal distribution
FitResult mf = s.MaximumLikelihoodFit((IList<double> p) => new NormalDistribution(p[0], p[1]), new double[] { mu + 1.0, sigma + 1.0 });
// it should find the parameters
Assert.IsTrue(mf.Dimension == 2);
Assert.IsTrue(mf.Parameter(0).ConfidenceInterval(0.99).ClosedContains(mu));
Assert.IsTrue(mf.Parameter(1).ConfidenceInterval(0.99).ClosedContains(sigma));
// now do our analytic fit
FitResult nf = NormalDistribution.FitToSample(s);
Assert.IsTrue(TestUtilities.IsNearlyEqual(mf.Parameter(0).Value, nf.Parameter(0).Value, 1.0E-4));
//Assert.IsTrue(TestUtilities.IsNearlyEqual(mf.Parameter(1).Value, nf.Parameter(1).Value, 1.0E-4));
Assert.IsTrue(TestUtilities.IsNearlyEqual(mf.Parameter(0).Uncertainty, nf.Parameter(0).Uncertainty, 1.0E-2));
Assert.IsTrue(TestUtilities.IsNearlyEqual(mf.Parameter(1).Uncertainty, nf.Parameter(1).Uncertainty, 1.0E-2));
}
public void BivariatePolynomialRegression()
{
// do a set of polynomial regression fits
// make sure not only that the fit parameters are what they should be, but that their variances/covariances are as claimed
Random rng = new Random(271828);
// define logistic parameters
double[] a = new double[] { 0.0, -1.0, 2.0, -3.0 };
// keep track of sample of returned a and b fit parameters
MultivariateSample A = new MultivariateSample(a.Length);
// also keep track of returned covariance estimates
// since these vary slightly from fit to fit, we will average them
SymmetricMatrix C = new SymmetricMatrix(a.Length);
// also keep track of test statistics
Sample F = new Sample();
// do 100 fits
for (int k = 0; k < 100; k++) {
// we should be able to draw x's from any distribution; noise should be drawn from a normal distribution
Distribution xd = new CauchyDistribution();
Distribution nd = new NormalDistribution(0.0, 4.0);
// generate a synthetic data set
BivariateSample s = new BivariateSample();
for (int j = 0; j < 20; j++) {
double x = xd.GetRandomValue(rng);
double y = nd.GetRandomValue(rng);
for (int i = 0; i < a.Length; i++) {
y += a[i] * MoreMath.Pow(x, i);
}
s.Add(x, y);
}
// do the regression
FitResult r = s.PolynomialRegression(a.Length - 1);
ColumnVector ps = r.Parameters;
//Console.WriteLine("{0} {1} {2}", ps[0], ps[1], ps[2]);
// record best fit parameters
A.Add(ps);
// record estimated covariances
C += r.CovarianceMatrix;
// record the fit statistic
F.Add(r.GoodnessOfFit.Statistic);
//Console.WriteLine("F={0}", r.GoodnessOfFit.Statistic);
}
C = (1.0 / A.Count) * C; // allow matrix division by real numbers
// check that mean parameter estimates are what they should be: the underlying population parameters
for (int i = 0; i < A.Dimension; i++) {
Console.WriteLine("{0} {1}", A.Column(i).PopulationMean, a[i]);
Assert.IsTrue(A.Column(i).PopulationMean.ConfidenceInterval(0.95).ClosedContains(a[i]));
}
// check that parameter covarainces are what they should be: the reported covariance estimates
for (int i = 0; i < A.Dimension; i++) {
for (int j = i; j < A.Dimension; j++) {
Console.WriteLine("{0} {1} {2} {3}", i, j, C[i, j], A.TwoColumns(i, j).PopulationCovariance);
Assert.IsTrue(A.TwoColumns(i, j).PopulationCovariance.ConfidenceInterval(0.95).ClosedContains(C[i, j]));
}
}
// check that F is distributed as it should be
//Console.WriteLine(fs.KolmogorovSmirnovTest(new FisherDistribution(2, 48)).LeftProbability);
}
public void NormalFit()
{
// pick mu >> sigma so that we get no negative values;
// otherwise the attempt to fit to an exponential will fail
Distribution distribution = new NormalDistribution(6.0, 2.0);
Sample sample = CreateSample(distribution, 100);
// fit to normal should be good
FitResult nfit = NormalDistribution.FitToSample(sample);
Console.WriteLine("P_n = {0}", nfit.GoodnessOfFit.LeftProbability);
Assert.IsTrue(nfit.GoodnessOfFit.LeftProbability < 0.95);
Assert.IsTrue(nfit.Parameter(0).ConfidenceInterval(0.95).ClosedContains(distribution.Mean));
Assert.IsTrue(nfit.Parameter(1).ConfidenceInterval(0.95).ClosedContains(distribution.StandardDeviation));
// fit to exponential should be bad
FitResult efit = ExponentialDistribution.FitToSample(sample);
Console.WriteLine("P_e = {0}", efit.GoodnessOfFit.LeftProbability);
Assert.IsTrue(efit.GoodnessOfFit.LeftProbability > 0.95);
}
public void BivariateLinearRegression()
{
// do a set of logistic regression fits
// make sure not only that the fit parameters are what they should be, but that their variances/covariances are as returned
Random rng = new Random(314159);
// define logistic parameters
double a0 = 2.0; double b0 = -1.0;
// keep track of sample of returned a and b fit parameters
BivariateSample ps = new BivariateSample();
// also keep track of returned covariance estimates
// since these vary slightly from fit to fit, we will average them
double caa = 0.0;
double cbb = 0.0;
double cab = 0.0;
// also keep track of test statistics
Sample fs = new Sample();
// do 100 fits
for (int k = 0; k < 100; k++) {
// we should be able to draw x's from any distribution; noise should be drawn from a normal distribution
Distribution xd = new LogisticDistribution();
Distribution nd = new NormalDistribution(0.0, 2.0);
// generate a synthetic data set
BivariateSample s = new BivariateSample();
for (int i = 0; i < 25; i++) {
double x = xd.GetRandomValue(rng);
double y = a0 + b0 * x + nd.GetRandomValue(rng);
s.Add(x, y);
}
// do the regression
FitResult r = s.LinearRegression();
// record best fit parameters
double a = r.Parameter(0).Value;
double b = r.Parameter(1).Value;
ps.Add(a, b);
// record estimated covariances
caa += r.Covariance(0, 0);
cbb += r.Covariance(1, 1);
cab += r.Covariance(0, 1);
// record the fit statistic
fs.Add(r.GoodnessOfFit.Statistic);
Console.WriteLine("F={0}", r.GoodnessOfFit.Statistic);
}
caa /= ps.Count;
cbb /= ps.Count;
cab /= ps.Count;
// check that mean parameter estimates are what they should be: the underlying population parameters
Assert.IsTrue(ps.X.PopulationMean.ConfidenceInterval(0.95).ClosedContains(a0));
Assert.IsTrue(ps.Y.PopulationMean.ConfidenceInterval(0.95).ClosedContains(b0));
Console.WriteLine("{0} {1}", caa, ps.X.PopulationVariance);
Console.WriteLine("{0} {1}", cbb, ps.Y.PopulationVariance);
// check that parameter covarainces are what they should be: the reported covariance estimates
Assert.IsTrue(ps.X.PopulationVariance.ConfidenceInterval(0.95).ClosedContains(caa));
Assert.IsTrue(ps.Y.PopulationVariance.ConfidenceInterval(0.95).ClosedContains(cbb));
Assert.IsTrue(ps.PopulationCovariance.ConfidenceInterval(0.95).ClosedContains(cab));
// check that F is distributed as it should be
Console.WriteLine(fs.KolmogorovSmirnovTest(new FisherDistribution(2, 48)).LeftProbability);
}
public void TimeNormalGenerators()
{
Random rng = new Random(1);
//IDeviateGenerator nRng = new BoxMullerNormalGenerator();
//IDeviateGenerator nRng = new PolarRejectionNormalDeviateGenerator();
//IDeviateGenerator nRng = new RatioOfUniformsNormalGenerator();
IDeviateGenerator nRng = new LevaNormalGenerator();
//Sample sample = new Sample();
Distribution nrm = new NormalDistribution();
Stopwatch timer = Stopwatch.StartNew();
double sum = 0.0;
for (int i = 0; i < 10000000; i++) {
sum += nrm.InverseLeftProbability(rng.NextDouble());
//sum += nRng.GetNext(rng);
//sample.Add(nRng.GetNext(rng));
}
timer.Stop();
//Console.WriteLine(sample.KolmogorovSmirnovTest(new NormalDistribution()).RightProbability);
Console.WriteLine(sum);
Console.WriteLine(timer.ElapsedMilliseconds);
}