📦 lumen

Title: Statistical Tests, Confidence Intervals, and Distributions
License: GPL (≥ 2)

🧩 Overview

lumen is the inferential layer of the DescToolsX ecosystem. It provides classical, robust, nonparametric and specialised hypothesis tests, confidence intervals for the common estimands, and a set of probability distributions that base R does not carry.

Tests return ordinary "htest" objects and intervals follow one argument convention throughout, so procedures from very different literatures can be combined without translating between interfaces.

The package is self-contained and does not require the higher-level packages of the suite.

📖 Documentation: https://andrisignorell.github.io/lumen/

⚙️ Installation

install.packages("lumen")

Or the development version from GitHub:

remotes::install_github("AndriSignorell/lumen")

📚 Core Features

🔹 Location Tests

• tTestA(), zTest(), yuenTTest(), signTest()
• brunnerMunzelTest(), siegelTukeyTest(), mosesTest()
• moodMedianTest(), vanWaerdenTest(), dscfTest()

🔹 Categorical Data

• barnardTest(), gTest(), cochranArmitageTest(), cochranQTest()
• breslowDayTest(), woolfTest(), mantelTrendTest()
• bhapkarTest(), stuartMaxwellTest(), lehmacherTest()

🔹 Goodness of Fit and Normality

• andersonDarlingTest(), cramerVonMisesTest(), lillieTest()
• shapiroFranciaTest(), jarqueBeraTest(), pearsonTest()
• hosmerLemeshowTest(), leCessieTest() — for logistic models

🔹 Multiple Comparisons and Post-hoc

• postHoc(), plot.PostHocTest()
• dunnTest(), dunnettTest(), conoverTest(), nemenyiTest()
• gamesHowellTest(), scheffeTest(), steelTest()

🔹 Trend, Randomness and Time Series

• adfTest(), kpssTest(), runsTest(), vonNeumannTest()
• bartelsRankTest(), coxStuartTest(), jonckheereTerpstraTest()
• pageTest(), durbinWatsonTest(), breuschGodfreyTest(), bpTest()

🔹 Confidence Intervals

• Proportions: binomCI(), binomDiffCI(), binomRatioCI(), multinomCI()
• Counts and rates: poissonCI(), poissonDiffCI(), poissonRatioCI()
• Location and scale: meanCI(), meanDiffCI(), medianCI(), quantileCI(), varCI(), sumCI()
• Correlation: corCI(), fisherZ()
• General: bootCI() for arbitrary statistics

🔹 Distributions

dpqr families for the Benford, Dirichlet, extreme value, Fréchet, GEV, GPD, Gompertz, Gumbel, reversed Gumbel, order-statistic, reverse Weibull and triangular distributions, plus cont.moments(), disc.moments() and extreme-value-moments().

🔹 Power and Sample Size

• meanCIn(), powerChisqTest(), sphericityEps(), scores()

🚀 Design Principles

• Consistent — lowerCamelCase API, uniform argument names and ordering across the whole DescToolsX suite
• Standard output — tests return "htest" objects and print like the ones in base R
• Fast — bootstrap and permutation routines implemented in C++ via Rcpp, RcppArmadillo and RcppParallel
• Documented — references to the original literature on every procedure

🧪 Example

library(lumen)

# a confidence interval for a proportion, several methods
binomCI(37, 100, method = "wilson")

# nonparametric alternative to the t test
brunnerMunzelTest(extra ~ group, data = sleep)

# post-hoc comparisons after an ANOVA
fit <- aov(breaks ~ tension, data = warpbreaks)
postHoc(fit, method = "hsd")

# goodness of fit
andersonDarlingTest(rnorm(100), null = "pnorm")

🧱 The Suite

lumen builds on bedrock (base utilities) and pharos (graphics). DescToolsX (descriptive statistics), alloy (modelling), pons (MS-Office) and swissValet (RStudio addins) complete the family.

🙏 Acknowledgements

Parts of the code and documentation were reviewed with the help of large language models (OpenAI Codex, Anthropic Claude). Every suggestion was assessed, edited and verified by the maintainer, who remains solely responsible for the content of this package.

📜 License

GPL (≥ 2)