distribution

The fit as an ordinary continuous distribution.

This is the object the whole procedure exists to produce: it has a density and a distribution function, it reports a mean and a variance, and it will supply a random variable that a simulation can draw from. Everything the library can already do with a fitted distribution, it can do with this — which is why the diagnostics below need no new statistics written for them.

A one-component candidate is returned as that component rather than as a mixture of one; see MixtureCandidate.toMixtureDistribution for why.