🌐 Website · 📚 Documentation
Probability distributions, descriptive statistics and statistical tests in pure Go (CGO=0), on go-ndarray.
32 univariate distributions (24 continuous, 8 discrete) and 3 multivariate ones, each with the methods of SciPy's frozen distributions — pdf, logpdf, cdf, logcdf, sf, logsf, ppf, isf, moments, entropy, interval, support, random draws and fits. Every distribution function is measured in ULPs against mpmath, with SciPy measured on the same points: over 1479 regions the error is at or below SciPy's in 1473, and the six exceptions are under 4.5 ulp (docs/accuracy.md). For the same seed, the draws of 17 distributions, Dirichlet and multinomial are bit-identical to SciPy's; the means, variances, the 13 quantile methods and the histograms are NumPy's bit for bit.
n := stats.Normal{Loc: 0, Scale: 1}
n.PDF(1.5) // 0.12951759566589172
n.LogCDF(-40) // -804.6084420137538
n.PPF(0.975) // 1.9599639845400538
// the same 1000 draws as scipy.stats.norm.rvs(10, 2, size=1000,
// random_state=np.random.default_rng(42)), to the bit
x, _ := stats.Normal{Loc: 10, Scale: 2}.Rvs(random.DefaultRNG(42), 1000)
r, _ := stats.TTestInd(a, b, &stats.TTestIndOptions{Welch: true})
qs, _ := stats.Percentiles(x, []float64{25, 50, 75}, nil)| Repo | What it is |
|---|---|
| stats | 24 continuous, 8 discrete and 3 multivariate distributions; descriptive statistics (NumPy's 13 quantile methods, weights, histograms, trimmed and binned statistics); hypothesis tests from t-tests and contingency tables to Kolmogorov–Smirnov, Shapiro–Wilk and Anderson–Darling; correlations and linear regression; Gaussian KDE and the bootstrap |
| docs | the documentation site (Hugo + Hextra), served at /docs/ |
| go-statistics.github.io | the landing page (Hugo) |
| brand | logo, icons and social images |
stats needs Go 1.27.2 or later, and its accuracy budgets are enforced on
amd64, arm64, 386, arm, riscv64, loong64, ppc64le and s390x. It is part of an
ecosystem with go-ndarray
(N-dimensional arrays and NumPy's random generator),
go-numerics (the special functions and
optimisers underneath), go-plot (figures),
go-dataframe (labelled tables)
and go-notebook (Go notebooks).