public void DistributionFunctionTest()
{
double[] ranks = { 1, 2, 3, 4, 5 };
int n1 = 2;
int n2 = 3;
int N = n1 + n2;
Assert.AreEqual(N, ranks.Length);
var target = new MannWhitneyDistribution(ranks, n1, n2);
// Number of possible combinations is 5!/(3!2!) = 10.
int nc = (int)Special.Binomial(5, 3);
Assert.AreEqual(10, nc);
double[] expected = { 0.0, 0.1, 0.1, 0.2, 0.2, 0.2, 0.1, 0.1, 0.0, 0.0 };
expected = expected.CumulativeSum();
for (int i = 0; i < expected.Length; i++)
{
// P(U<=i)
double actual = target.DistributionFunction(i);
Assert.AreEqual(expected[i], actual,1e-10);
}
}
/// <summary>
/// Gets the cumulative distribution function (cdf) for
/// this distribution evaluated at point <c>k</c>.
/// </summary>
///
/// <param name="k">A single point in the distribution range.</param>
///
/// <remarks>
/// The Cumulative Distribution Function (CDF) describes the cumulative
/// probability that a given value or any value smaller than it will occur.
/// </remarks>
///
public override double DistributionFunction(int k)
{
if (k < Math.Max(0, n + m - N))
{
return(0);
}
if (k > Math.Min(m, n))
{
k = Math.Min(m, n);
}
// This is a really naive implementation. A better approach
// is described in (Trong Wu; An accurate computation of the
// hypergeometric distribution function, 1993)
double sum = 0;
for (int i = 0; i <= k; i++)
{
sum += (Special.Binomial(m, i) * Special.Binomial(N - m, n - i));
}
return(sum / Special.Binomial(N, n));
}
public void ProbabilityDensityFunctionTest()
{
double[] ranks = { 1, 2, 3, 4, 5 };
int n1 = 2;
int n2 = 3;
int N = n1 + n2;
Assert.AreEqual(N, ranks.Length);
var target = new MannWhitneyDistribution(ranks, n1, n2);
// Number of possible combinations is 5!/(3!2!) = 10.
int nc = (int)Special.Binomial(5, 3);
Assert.AreEqual(10, nc);
double[] expected = { 0.1, 0.1, 0.2, 0.2, 0.2, 0.1, 0.1, 0.0, 0.0 };
double sum = 0;
for (int i = 0; i < expected.Length; i++)
{
// P(U=i)
double actual = target.ProbabilityDensityFunction(i);
Assert.AreEqual(expected[i], actual);
sum += actual;
}
Assert.AreEqual(1, sum);
}
/// <summary>
/// Constructs a Mann-Whitney's U-statistic distribution.
/// </summary>
///
/// <param name="ranks">The rank statistics.</param>
/// <param name="n1">The number of observations in the first sample.</param>
/// <param name="n2">The number of observations in the second sample.</param>
///
public MannWhitneyDistribution(double[] ranks, int n1, int n2)
{
this.Ranks = ranks;
this.Samples1 = n1;
this.Samples2 = n2;
int nt = n1 + n2;
this.smallSample = (n1 <= 30 && n2 <= 30);
if (smallSample)
{
// For a small sample (< 30) the distribution is exact.
int nc = (int)Special.Binomial(nt, n1);
table = new double[nc];
int i = 0; // Consider all possible combinations of samples
foreach (double[] combination in Combinatorics.Combinations(Ranks, n1))
{
table[i++] = USample1(combination, Samples2);
}
Array.Sort(table);
}
}
/// <summary>
/// Gets the probability mass function (pmf) for
/// this distribution evaluated at point <c>x</c>.
/// </summary>
///
/// <param name="k">A single point in the distribution range.</param>
///
/// <returns>
/// The probability of <c>k</c> occurring
/// in the current distribution.
/// </returns>
///
/// <remarks>
/// The Probability Mass Function (PMF) describes the
/// probability that a given value <c>x</c> will occur.
/// </remarks>
///
protected internal override double InnerProbabilityMassFunction(int k)
{
double a = Special.Binomial(m, k);
double b = Special.Binomial(N - m, n - k);
double c = Special.Binomial(N, n);
return((a * b) / c);
}
/// <summary>
/// Gets the probability mass function (pmf) for
/// this distribution evaluated at point <c>x</c>.
/// </summary>
///
/// <param name="k">A single point in the distribution range.</param>
///
/// <returns>
/// The probability of <c>k</c> occurring
/// in the current distribution.
/// </returns>
///
/// <remarks>
/// The Probability Mass Function (PMF) describes the
/// probability that a given value <c>x</c> will occur.
/// </remarks>
///
public override double ProbabilityMassFunction(int k)
{
if (k < 0)
{
return(0);
}
return(Special.Binomial(k + r - 1, r - 1) * Math.Pow(1 - p, k) * Math.Pow(p, r));
}
/// <summary>
/// Gets the probability mass function (pmf) for
/// this distribution evaluated at point <c>x</c>.
/// </summary>
///
/// <param name="k">A single point in the distribution range.</param>
///
/// <returns>
/// The probability of <c>k</c> occurring
/// in the current distribution.
/// </returns>
///
/// <remarks>
/// The Probability Mass Function (PMF) describes the
/// probability that a given value <c>x</c> will occur.
/// </remarks>
///
public override double ProbabilityMassFunction(int k)
{
if (k < 0 || k > numberOfTrials)
{
return(0);
}
return(Special.Binomial(numberOfTrials, k) * Math.Pow(probability, k) * Math.Pow(1 - probability, numberOfTrials - k));
}
/// <summary>
/// Gets the probability mass function (pmf) for
/// this distribution evaluated at point <c>x</c>.
/// </summary>
///
/// <param name="k">A single point in the distribution range.</param>
///
/// <returns>
/// The probability of <c>k</c> occurring
/// in the current distribution.
/// </returns>
///
/// <remarks>
/// The Probability Mass Function (PMF) describes the
/// probability that a given value <c>x</c> will occur.
/// </remarks>
///
public override double ProbabilityMassFunction(int k)
{
if (k < Math.Max(0, n + m - N) || k > Math.Min(m, n))
{
return(0);
}
return((Special.Binomial(m, k) * Special.Binomial(N - m, n - k))
/ Special.Binomial(N, n));
}
/// <summary>
/// Gets the probability mass function (pmf) for
/// this distribution evaluated at point <c>x</c>.
/// </summary>
///
/// <param name="k">A single point in the distribution range.</param>
///
/// <returns>
/// The probability of <c>k</c> occurring
/// in the current distribution.
/// </returns>
///
/// <remarks>
/// The Probability Mass Function (PMF) describes the
/// probability that a given value <c>x</c> will occur.
/// </remarks>
///
public override double ProbabilityMassFunction(int k)
{
if (k < Math.Max(0, n + m - N) || k > Math.Min(m, n))
{
return(0);
}
double a = Special.Binomial(m, k);
double b = Special.Binomial(N - m, n - k);
double c = Special.Binomial(N, n);
return((a * b) / c);
}
/// <summary>
/// Returns the value of the probability density function.
/// </summary>
/// <param name="x">Value</param>
/// <returns>float precision floating point number</returns>
public float Function(float x)
{
if (x < Math.Max(0, k + d - n) || x > Math.Min(d, k))
{
return(0);
}
float a = Special.Binomial(d, x);
float b = Special.Binomial(n - d, k - x);
float c = Special.Binomial(n, k);
return((a * b) / c);
}
public void BinomialTest2()
{
int n = 6;
int k = 4;
double expected = 15;
double actual = Special.Binomial(n, k);
Assert.AreEqual(expected, actual);
n = 100;
k = 47;
expected = 8.441348728306404e+28;
actual = Special.Binomial(n, k);
Assert.AreEqual(expected, actual, 1e+16);
}
public void BinomialTest()
{
int n = 63;
int k = 6;
double expected = 67945521;
double actual;
actual = Special.Binomial(n, k);
Assert.AreEqual(expected, actual);
n = 42;
k = 12;
expected = 11058116888;
actual = Special.Binomial(n, k);
Assert.AreEqual(expected, actual);
}
/// <summary>
/// Constructs a Mann-Whitney's U-statistic distribution.
/// </summary>
///
/// <param name="ranks">The rank statistics.</param>
/// <param name="n1">The number of observations in the first sample.</param>
/// <param name="n2">The number of observations in the second sample.</param>
///
public MannWhitneyDistribution(double[] ranks,
[PositiveInteger] int n1, [PositiveInteger] int n2)
{
this.Ranks = ranks;
this.Samples1 = n1;
this.Samples2 = n2;
int nt = n1 + n2;
if (n1 <= 0)
{
throw new ArgumentOutOfRangeException("n1",
"The first number of samples must be positive.");
}
if (n2 <= 0)
{
throw new ArgumentOutOfRangeException("n2",
"The second number of samples must be positive.");
}
this.smallSample = (n1 <= 30 && n2 <= 30);
if (smallSample)
{
// For a small sample (< 30) the distribution is exact.
int nc = (int)Special.Binomial(nt, n1);
table = new double[nc];
int i = 0; // Consider all possible combinations of samples
foreach (double[] combination in Combinatorics.Combinations(Ranks, n1))
{
table[i++] = USample1(combination, Samples2);
}
Array.Sort(table);
}
}
private void initExactMethod(double[] ranks)
{
int min = Math.Min(n1, n2);
long combinations = (long)Special.Binomial(n1 + n2, min);
this.table = new double[combinations];
#if NET35
var seq = EnumerableEx.Zip <double[], long, Tuple <double[], long> >(
Combinatorics.Combinations(ranks, min), Vector.Range(combinations),
(double[] c, long i) => new Tuple <double[], long>(c, i));
#else
var seq = Enumerable.Zip <double[], long, Tuple <double[], long> >(
Combinatorics.Combinations(ranks, min), Vector.Range(combinations),
(double[] c, long i) => new Tuple <double[], long>(c, i));
#endif
Parallel.ForEach(seq, i =>
{
this.table[i.Item2] = MannWhitneyU(i.Item1);
});
Array.Sort(table);
}
protected override void EndProcessing()
{
WriteObject(Special.Binomial(N, K));
}
/// <summary>
/// Gets the probability mass function (pmf) for
/// this distribution evaluated at point <c>x</c>.
/// </summary>
///
/// <param name="k">A single point in the distribution range.</param>
///
/// <returns>
/// The probability of <c>k</c> occurring
/// in the current distribution.
/// </returns>
///
/// <remarks>
/// The Probability Mass Function (PMF) describes the
/// probability that a given value <c>x</c> will occur.
/// </remarks>
///
protected internal override double InnerProbabilityMassFunction(int k)
{
return(Special.Binomial(k + r - 1, r - 1) * Math.Pow(1 - p, k) * Math.Pow(p, r));
}