public void LinearLogisticRegressionSimple()
{
Polynomial m = Polynomial.FromCoefficients(-1.0, 2.0);
FrameTable table = new FrameTable();
table.AddColumn <double>("x");
table.AddColumn <string>("z");
Random rng = new Random(2);
ContinuousDistribution xDistribution = new CauchyDistribution(4.0, 2.0);
for (int i = 0; i < 24; i++)
{
double x = xDistribution.GetRandomValue(rng);
double y = m.Evaluate(x);
double p = 1.0 / (1.0 + Math.Exp(-y));
bool z = (rng.NextDouble() < p);
table.AddRow(x, z.ToString());
}
LinearLogisticRegressionResult fit = table["z"].As((string s) => Boolean.Parse(s)).LinearLogisticRegression(table["x"].As <double>());
Assert.IsTrue(fit.Intercept.ConfidenceInterval(0.99).ClosedContains(m.Coefficient(0)));
Assert.IsTrue(fit.Slope.ConfidenceInterval(0.99).ClosedContains(m.Coefficient(1)));
}
public void MultivariateLinearRegressionVariances()
{
// define model y = a + b0 * x0 + b1 * x1 + noise
double a = -3.0;
double b0 = 2.0;
double b1 = -1.0;
ContinuousDistribution x0distribution = new LaplaceDistribution();
ContinuousDistribution x1distribution = new CauchyDistribution();
ContinuousDistribution eDistribution = new NormalDistribution(0.0, 4.0);
FrameTable data = new FrameTable();
data.AddColumns <double>("a", "da", "b0", "db0", "b1", "db1", "ab1Cov", "p", "dp");
// draw a sample from the model
Random rng = new Random(4);
for (int j = 0; j < 64; j++)
{
List <double> x0s = new List <double>();
List <double> x1s = new List <double>();
List <double> ys = new List <double>();
for (int i = 0; i < 16; i++)
{
double x0 = x0distribution.GetRandomValue(rng);
double x1 = x1distribution.GetRandomValue(rng);
double e = eDistribution.GetRandomValue(rng);
double y = a + b0 * x0 + b1 * x1 + e;
x0s.Add(x0);
x1s.Add(x1);
ys.Add(y);
}
// do a linear regression fit on the model
MultiLinearRegressionResult result = ys.MultiLinearRegression(
new Dictionary <string, IReadOnlyList <double> > {
{ "x0", x0s }, { "x1", x1s }
}
);
UncertainValue pp = result.Predict(-5.0, 6.0);
data.AddRow(
result.Intercept.Value, result.Intercept.Uncertainty,
result.CoefficientOf("x0").Value, result.CoefficientOf("x0").Uncertainty,
result.CoefficientOf("x1").Value, result.CoefficientOf("x1").Uncertainty,
result.Parameters.CovarianceOf("Intercept", "x1"),
pp.Value, pp.Uncertainty
);
}
// The estimated parameters should agree with the model that generated the data.
// The variances of the estimates should agree with the claimed variances
Assert.IsTrue(data["a"].As <double>().PopulationStandardDeviation().ConfidenceInterval(0.99).ClosedContains(data["da"].As <double>().Mean()));
Assert.IsTrue(data["b0"].As <double>().PopulationStandardDeviation().ConfidenceInterval(0.99).ClosedContains(data["db0"].As <double>().Mean()));
Assert.IsTrue(data["b1"].As <double>().PopulationStandardDeviation().ConfidenceInterval(0.99).ClosedContains(data["db1"].As <double>().Mean()));
Assert.IsTrue(data["a"].As <double>().PopulationCovariance(data["b1"].As <double>()).ConfidenceInterval(0.99).ClosedContains(data["ab1Cov"].As <double>().Mean()));
Assert.IsTrue(data["p"].As <double>().PopulationStandardDeviation().ConfidenceInterval(0.99).ClosedContains(data["dp"].As <double>().Median()));
}
public void SpearmanNullDistributionTest()
{
// Pick independent distributions for x and y, which needn't be normal and needn't be related.
ContinuousDistribution xDistrubtion = new UniformDistribution();
ContinuousDistribution yDistribution = new CauchyDistribution();
Random rng = new Random(1);
// Generate bivariate samples of various sizes
foreach (int n in TestUtilities.GenerateIntegerValues(4, 64, 8))
{
Sample testStatistics = new Sample();
ContinuousDistribution testDistribution = null;
for (int i = 0; i < 128; i++)
{
BivariateSample sample = new BivariateSample();
for (int j = 0; j < n; j++)
{
sample.Add(xDistrubtion.GetRandomValue(rng), yDistribution.GetRandomValue(rng));
}
TestResult result = sample.SpearmanRhoTest();
testStatistics.Add(result.Statistic);
testDistribution = result.Distribution;
}
TestResult r2 = testStatistics.KolmogorovSmirnovTest(testDistribution);
Assert.IsTrue(r2.Probability > 0.05);
Assert.IsTrue(testStatistics.PopulationMean.ConfidenceInterval(0.99).ClosedContains(testDistribution.Mean));
Assert.IsTrue(testStatistics.PopulationVariance.ConfidenceInterval(0.99).ClosedContains(testDistribution.Variance));
}
}
public void MultivariateLinearRegressionSimple()
{
// define model y = a + b0 * x0 + b1 * x1 + noise
double a = 1.0;
double b0 = -2.0;
double b1 = 3.0;
ContinuousDistribution x0distribution = new CauchyDistribution(10.0, 5.0);
ContinuousDistribution x1distribution = new UniformDistribution(Interval.FromEndpoints(-10.0, 20.0));
ContinuousDistribution noise = new NormalDistribution(0.0, 10.0);
// draw a sample from the model
Random rng = new Random(1);
MultivariateSample sample = new MultivariateSample("x0", "x1", "y");
FrameTable table = new FrameTable();
table.AddColumns <double>("x0", "x1", "y");
for (int i = 0; i < 100; i++)
{
double x0 = x0distribution.GetRandomValue(rng);
double x1 = x1distribution.GetRandomValue(rng);
double eps = noise.GetRandomValue(rng);
double y = a + b0 * x0 + b1 * x1 + eps;
sample.Add(x0, x1, y);
table.AddRow(x0, x1, y);
}
// do a linear regression fit on the model
ParameterCollection oldResult = sample.LinearRegression(2).Parameters;
MultiLinearRegressionResult newResult = table["y"].As <double>().MultiLinearRegression(
table["x0"].As <double>(), table["x1"].As <double>()
);
// the result should have the appropriate dimension
Assert.IsTrue(oldResult.Count == 3);
Assert.IsTrue(newResult.Parameters.Count == 3);
// The parameters should match the model
Assert.IsTrue(oldResult[0].Estimate.ConfidenceInterval(0.90).ClosedContains(b0));
Assert.IsTrue(oldResult[1].Estimate.ConfidenceInterval(0.90).ClosedContains(b1));
Assert.IsTrue(oldResult[2].Estimate.ConfidenceInterval(0.90).ClosedContains(a));
Assert.IsTrue(newResult.CoefficientOf(0).ConfidenceInterval(0.99).ClosedContains(b0));
Assert.IsTrue(newResult.CoefficientOf("x1").ConfidenceInterval(0.99).ClosedContains(b1));
Assert.IsTrue(newResult.Intercept.ConfidenceInterval(0.99).ClosedContains(a));
// The residuals should be compatible with the model predictions
for (int i = 0; i < table.Rows.Count; i++)
{
FrameRow row = table.Rows[i];
double x0 = (double)row["x0"];
double x1 = (double)row["x1"];
double yp = newResult.Predict(x0, x1).Value;
double y = (double)row["y"];
Assert.IsTrue(TestUtilities.IsNearlyEqual(newResult.Residuals[i], y - yp));
}
}
public void BivariateAssociationDiscreteNullDistribution()
{
Random rng = new Random(1);
// Pick very non-normal distributions for our non-parameteric tests
ContinuousDistribution xd = new FrechetDistribution(1.0);
ContinuousDistribution yd = new CauchyDistribution();
// Pick small sample sizes to get exact distributions
foreach (int n in TestUtilities.GenerateIntegerValues(4, 24, 4))
{
// Do a bunch of test runs, recording reported statistic for each.
List <int> spearmanStatistics = new List <int>();
List <int> kendallStatistics = new List <int>();
DiscreteDistribution spearmanDistribution = null;
DiscreteDistribution kendallDistribution = null;
for (int i = 0; i < 512; i++)
{
List <double> x = new List <double>();
List <double> y = new List <double>();
for (int j = 0; j < n; j++)
{
x.Add(xd.GetRandomValue(rng));
y.Add(yd.GetRandomValue(rng));
}
DiscreteTestStatistic spearman = Bivariate.SpearmanRhoTest(x, y).UnderlyingStatistic;
if (spearman != null)
{
spearmanStatistics.Add(spearman.Value);
spearmanDistribution = spearman.Distribution;
}
DiscreteTestStatistic kendall = Bivariate.KendallTauTest(x, y).UnderlyingStatistic;
if (kendall != null)
{
kendallStatistics.Add(kendall.Value);
kendallDistribution = kendall.Distribution;
}
}
// Test whether statistics are actually distributed as claimed
if (spearmanDistribution != null)
{
TestResult spearmanChiSquared = spearmanStatistics.ChiSquaredTest(spearmanDistribution);
Assert.IsTrue(spearmanChiSquared.Probability > 0.01);
}
if (kendallDistribution != null)
{
TestResult kendallChiSquared = kendallStatistics.ChiSquaredTest(kendallDistribution);
Assert.IsTrue(kendallChiSquared.Probability > 0.01);
}
}
}
public void BivariateNullAssociation()
{
Random rng = new Random(31415926);
// Create a data structure to hold the results of Pearson, Spearman, and Kendall tests.
FrameTable data = new FrameTable();
data.AddColumn <double>("r");
data.AddColumn <double>("ρ");
data.AddColumn <double>("τ");
// Create variables to hold the claimed distribution of each test statistic.
ContinuousDistribution PRD = null;
ContinuousDistribution SRD = null;
ContinuousDistribution KTD = null;
// Generate a large number of bivariate samples and conduct our three tests on each.
ContinuousDistribution xDistribution = new LognormalDistribution();
ContinuousDistribution yDistribution = new CauchyDistribution();
for (int j = 0; j < 100; j++)
{
List <double> x = new List <double>();
List <double> y = new List <double>();
for (int i = 0; i < 100; i++)
{
x.Add(xDistribution.GetRandomValue(rng));
y.Add(yDistribution.GetRandomValue(rng));
}
TestResult PR = Bivariate.PearsonRTest(x, y);
TestResult SR = Bivariate.SpearmanRhoTest(x, y);
TestResult KT = Bivariate.KendallTauTest(x, y);
PRD = PR.Statistic.Distribution;
SRD = SR.Statistic.Distribution;
KTD = KT.Statistic.Distribution;
data.AddRow(new Dictionary <string, object>()
{
{ "r", PR.Statistic.Value }, { "ρ", SR.Statistic.Value }, { "τ", KT.Statistic.Value }
});
}
Assert.IsTrue(data["r"].As <double>().KolmogorovSmirnovTest(PRD).Probability > 0.05);
Assert.IsTrue(data["ρ"].As <double>().KolmogorovSmirnovTest(SRD).Probability > 0.05);
Assert.IsTrue(data["τ"].As <double>().KolmogorovSmirnovTest(KTD).Probability > 0.05);
}
public void BivariatePolynomialRegressionCovariance()
{
// 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
ContinuousDistribution xd = new CauchyDistribution();
ContinuousDistribution 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
PolynomialRegressionResult r = s.PolynomialRegression(a.Length - 1);
ColumnVector ps = r.Parameters.ValuesVector;
// record best fit parameters
A.Add(ps);
// record estimated covariances
C += r.Parameters.CovarianceMatrix;
// record the fit statistic
F.Add(r.F.Statistic.Value);
}
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++)
{
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++)
{
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 LinearRegressionSimple()
{
double a = -1.0;
double b = 2.0;
ContinuousDistribution xDistribution = new CauchyDistribution();
ContinuousDistribution eDistribution = new NormalDistribution();
int n = 16;
Random rng = new Random(1);
double[] x = new double[n];
double[] y = new double[n];
for (int i = 0; i < 16; i++)
{
x[i] = xDistribution.GetRandomValue(rng);
y[i] = a + b * x[i] + eDistribution.GetRandomValue(rng);
}
LinearRegressionResult result = y.LinearRegression(x);
// Parameters should be right
Assert.IsTrue(result.Intercept.ConfidenceInterval(0.95).ClosedContains(a));
Assert.IsTrue(result.Slope.ConfidenceInterval(0.95).ClosedContains(b));
// Reported values should be consistent
Assert.IsTrue(result.Intercept == result.Parameters["Intercept"].Estimate);
Assert.IsTrue(result.Intercept.Value == result.Parameters.ValuesVector[result.Parameters.IndexOf("Intercept")]);
Assert.IsTrue(TestUtilities.IsNearlyEqual(result.Intercept.Uncertainty, Math.Sqrt(result.Parameters.VarianceOf("Intercept"))));
Assert.IsTrue(result.Slope == result.Parameters["Slope"].Estimate);
Assert.IsTrue(result.Slope.Value == result.Parameters.ValuesVector[result.Parameters.IndexOf("Slope")]);
Assert.IsTrue(TestUtilities.IsNearlyEqual(result.Slope.Uncertainty, Math.Sqrt(result.Parameters.VarianceOf("Slope"))));
// Residuals should agree with definition
for (int i = 0; i < x.Length; i++)
{
double yp = result.Predict(x[i]).Value;
Assert.IsTrue(TestUtilities.IsNearlyEqual(result.Residuals[i], y[i] - yp));
}
// R and R-squared agree
Assert.IsTrue(TestUtilities.IsNearlyEqual(result.RSquared, MoreMath.Sqr(result.R.Statistic.Value)));
// F-test and R-test agree
Assert.IsTrue(TestUtilities.IsNearlyEqual(result.F.Probability, result.R.Probability));
// ANOVA's sums of squares are correct
double SST = y.Variance() * y.Length;
Assert.IsTrue(TestUtilities.IsNearlyEqual(SST, result.Anova.Total.SumOfSquares));
double SSR = 0.0;
foreach (double z in result.Residuals)
{
SSR += z * z;
}
Assert.IsTrue(TestUtilities.IsNearlyEqual(SSR, result.Anova.Residual.SumOfSquares));
Assert.IsTrue(TestUtilities.IsNearlyEqual(SSR, result.SumOfSquaredResiduals));
// R is same as correlation coefficient
Assert.IsTrue(TestUtilities.IsNearlyEqual(x.CorrelationCoefficient(y), result.R.Statistic.Value));
}