@JuliaRegistrator register

Release notes:

First release; basic functionality including constructing zero-inflated non-negative likelihoods, calculating the log-likelihood of observations, and generating random samples.

These are actually distributions, with e.g. mean etc. that can be calculated based on the underlying distributions. Might take some thought about how to deal with discrete zero probability vs continuous density.

Methods to implement:

• rand(::AbstractRNG, d::UnivariateDistribution)
• sampler(d::Distribution)
• pdf(d::UnivariateDistribution, x::Real)
• logpdf(d::UnivariateDistribution, x::Real)
• cdf(d::UnivariateDistribution, x::Real)
• quantile(d::UnivariateDistribution, q::Real)
• minimum(d::UnivariateDistribution)
• maximum(d::UnivariateDistribution)
• insupport(d::UnivariateDistribution, x::Real)

It is also recommended that one also implements the following statistics functions:

• mean(d::UnivariateDistribution)
• var(d::UnivariateDistribution)
• modes(d::UnivariateDistribution)
• mode(d::UnivariateDistribution)
• skewness(d::UnivariateDistribution)
• kurtosis(d::Distribution, ::Bool)
• entropy(d::UnivariateDistribution, ::Real)
• mgf(d::UnivariateDistribution, ::Any)
• cf(d::UnivariateDistribution, ::Any)

@JuliaRegistrator register