@danielsimonjr/mathts-statistics

Statistics & probability for MathTS — a focused, standalone library covering descriptive statistics, probability distributions, hypothesis tests, and combinatorics.

Every name this package exports is implemented in @danielsimonjr/mathts-functions. This package does not define a second copy of a distribution.

The implementations live in @danielsimonjr/mathts-functions and are re-exported here, not duplicated — this package is a curated entry point over the statistics/probability surface. Every function is externally oracle-pinned (SciPy / NumPy / mpmath / closed form) in the functions package's test suite, so the numbers are trustworthy, not just green.

Install

npm install @danielsimonjr/mathts-statistics

What's included

• Descriptive statistics — mean, median, mode, variance, std, quantileSeq, mad (median absolute deviation), iqr, sem, zscore, moment, skewness, kurtosis, gmean, hmean, cov, corr, corrcoef, rankdata, ptp, trimmedMean, variation, describe, logsumexp, softmax, detrend, movingAverage, ewma, acf, and the correlation helpers pearsonr, spearman, spearmanr, kendalltau / kendallTau.
• Parallel reductions — parallelStat* helpers mirror the typed reductions on Float64Array; most op thresholds in ComputePool are 'never' today (inline JS), but the API stays async-friendly where the typed layer returns a Promise. Factory mad is median absolute deviation; parallelStatMAD / parallelStatMeanAD are mean absolute deviation from the mean (not the same statistic). quickSelect / medianSelect / minSelect / maxSelect are synchronous selection utilities.
• Regression & resampling — linearRegression, polyFit, linregress, pearsonr, spearman, spearmanr, kendalltau, meanCI, proportionCI, bootstrapCI, permutationTest, mahalanobis, hotellingT2, kmeans, spectralClustering.
• Probability distributions — distribution objects with .pdf/.cdf/.quantile/.mean/.variance/.sample (normalDist, gammaDist, betaDist, binomialDist, poissonDist, exponentialDist, chiSquaredDist, fDist, tDist, logNormalDist, uniformDist, weibullDist, hypergeometricDist, negativeBinomialDist, paretoDist, rayleighDist, triangularDist, discreteUniformDist, gumbelDist, invGaussDist, multivariateNormal), plus free CDF/quantile functions, studentizedRangeCDF / studentizedRangeQuantile, and entropy / jsDivergence.
• Hypothesis tests — studentTTest (one/two-sample), studentTTestPaired, anova, anova2, chiSquareTest, kolmogorovSmirnovTest + kolmogorovSmirnov2Test, mannWhitneyTest, shapiroWilkTest, leveneTest, bartlettTest, proportionZTest, binomialTest, fTest, jarqueBera, kruskalWallis, wilcoxon, fisherExact, tukeyHSD, andersonDarlingTest, dagostinoTest, friedmanTest, multipleComparison, principalComponentAnalysis.
• Probability & combinatorics — combinations, combinationsWithRep, permutations, multinomial, factorial, doubleFactorial, risingFactorial, fallingFactorial, subfactorial, bernoulli, gamma, lgamma, kldivergence, and seedable RNG (random, randomInt, pickRandom, seedProbabilityRng).

Example

import {
  anova,
  kendallTau,
  hypergeometricDist,
  kolmogorovSmirnov2Test,
} from '@danielsimonjr/mathts-statistics';

anova([
  [8.1, 8.3, 7.9],
  [9.1, 9.5, 8.9],
]); // one-way ANOVA F + p
kendallTau([1, 2, 3, 4, 5], [2, 1, 4, 3, 5]); // 0.6
hypergeometricDist(50, 5, 10).pdf(1); // 0.4313371972
kolmogorovSmirnov2Test([1, 2, 3], [2, 3, 4]); // { statistic, pValue }

License

MIT