Stats Lodestar.Stats

Development build. This page describes main, not a released package. The latest published Lodestar.Stats is 0.5.0 — read its documentation.

Home › Stats

Lodestar.Stats

Classical hypothesis tests with scipy.stats' semantics: Student and Welch t, Mann-Whitney U, Wilcoxon signed-rank, chi-square and its contingency form, Fisher exact, Kolmogorov-Smirnov, Shapiro-Wilk, Anderson-Darling, one-way ANOVA, Kruskal-Wallis, Levene, Bartlett, Friedman, the binomial test and the correlation tests. Each returns its statistic and p-value, with scipy's alternatives and nan_policy. It also publishes the four distribution tails its neighbours need.

Install

dotnet add package Lodestar.Stats

Example

using Lodestar.Stats;

double[] before = [102.0, 98.0, 110.0, 105.0, 99.0];
double[] after = [95.0, 92.0, 99.0, 91.0, 97.0];

TTestResult result = TTest.Independent(before, after);   // Welch's test by default
bool significant = result.PValue < 0.05;                // True

Parity

Replayed against scipy.stats. docs/equivalence.md maps each Python call to its C# counterpart, with every deliberate divergence.

Dependencies

A core package (decision 0003), built for net10.0 and netstandard2.0:

• Lodestar.Abstractions 0.2.0 or later

Documentation

• Guide: hypothesis testing
• Reference: stats/nanpolicy
• Reference: stats/tails
• Reference: stats/tests
• Changelog
• Performance
• All packages