mdists

An R package providing a collection of uncommon or reparameterized probability distributions. Many standard distributions (beta, gamma, log-normal) are wrapped with mean/standard-deviation parameterizations for convenience.

Installation

devtools::install_github("bamonroe/mdists")

Distributions

Mean-parameterized distributions

These let you specify distributions by their mean and standard deviation instead of shape parameters:

Function	Description
mdbeta, mpbeta, mqbeta, mrbeta	Beta (d/p/q/r) via mean and sd
mdgamma, mpgamma, mqgamma, mrgamma	Gamma (d/p/q/r) via mean and sd
mdlnorm, mplnorm, mqlnorm, mrlnorm	Log-normal (d/p/q/r) via mean and sd

Beta-binomial

Function	Description
bbinom	Beta-binomial PMF
bbinom2	Beta-binomial PMF (numerically stable formulation)

Binomial variants

Function	Description
dbinom2	Linear mixture of two binomial distributions
sbinom	Binomial with a shape/scale parameter

Utilities

Function	Description
lnorm	Convert normal-space mean/sd to log-space parameters

Example

library(mdists)

# Beta distribution specified by mean and sd
mdbeta(0.5, mu = 0.3, sigma = 0.1)

# Gamma random draws by mean and sd
mrgamma(100, mu = 5, sigma = 2)

# Beta-binomial probability
bbinom(n = 10, k = 3, a = 2, b = 5)

License

GPL-3