Stats 0.2.0 distributions - CyrilB1531/lodestar GitHub Wiki

Lodestar.Stats 0.2.0. This page is frozen at that release. Read the current documentation for what main says now. A link to a decision or a migration page follows main, and leaves the archive.

Distributions

The tail probabilities and the quantile a caller holding its own statistic needs.

public static class Distributions

Example — a coefficient's two-sided p-value and its 95% multiplier, the two numbers a regression table prints beside an estimate.

using Lodestar.Stats;

// A t of 2.0 on 12 residual degrees of freedom.
double twoSided = 2.0 * Distributions.StudentSf(2.0, 12.0);   // => 0.0686550…
double multiplier = Distributions.StudentQuantile(0.975, 12.0); // => 2.1788128296672298

// And the overall F test of a model with two regressors and twenty residual df.
double overall = Distributions.FisherSf(4.0, 2.0, 20.0);        // => 0.0345716…

Remarks — these are the four members Lodestar.Stats.Regression needs, and no more. The machinery underneath — log-gamma, the incomplete beta and gamma, the normal tail — stays internal; decisions/0081 says why, and the index page says what publishing these four cost.

Applies to — net10.0, netstandard2.0.

See alsoTTest, OneWayAnova, the distributions index.

Members

Member What it does
Distributions.FisherSf The upper tail of the F distribution.
Distributions.StudentQuantile The value a Student's t falls below with a given probability.
Distributions.StudentSf The upper tail of Student's t.