Distributions.jl

Distributions.jl copied to clipboard

Von mises range fix

3

This pull request reverts part of #965 which added bound checking to the pdf of several distributions. This shouldn't be done for the von Mises distribution as it is designed...

Astro-mh

fix some bugs related to the Chernoff distribution.

9

Some functionality of the Chernoff() distribution depended on the newton() function which was not in the dependent list. I added the dependency and also polish some codes.

sl-solution

Re-export from LinearAlgebra and FillArrays

7

Deprecations introduced in #1362 suggest alternatives that use the following: - `LinearAlgebra.Diagonal` - `LinearAlgebra.I` - `FillArrays.Fill` It seems like these should then also be re-exported, in particular since Distributions.jl is...

torfjelde

``` cdf(Chernoff(), -0.5) === 1-cdf(Chernoff(), 0.5) ``` gives false as ``` cdf(Chernoff(), -0.5) + cdf(Chernoff(), 0.5) ~ 1 ``` Should we change the [documentation](https://github.com/JuliaStats/Distributions.jl/blob/ed52948fb47e7e0aecaf81e99ebb090c2738cf0e/src/univariate/continuous/chernoff.jl#L44)?

dnabanita7

Add documentation about the algorithm behind computing MvNormal samples.

3

In my very humble opinion, it would be useful to add some information about which algorithm is used for the generation of `MvNormal` samples. For instance, is Cholesky decomposition used...

ignace-computing

Generic truncated normal & exponential

10

Fixes https://github.com/JuliaStats/Distributions.jl/issues/1089. Since the sampler for truncated normal takes samples from `Exponential`, I also modified the `Exponential` sampler to return values following the parameter type.

cossio

incongruency in `params` of MvNormalCanon vs NormalCanon

1

In contrast to the (co)variance vs standard deviation inconsistency between Normal and MvNormal (https://github.com/JuliaStats/Distributions.jl/issues/584), NormalCanon and MvNormalCanon both are defined in terms of the natural parameters, \eta = \mu *...

st--

Undeprecate quantile() for arrays

17

Fixes #1150 by undeprecating omission of the broadcast operator `.` when using `quantile`, `quantile!`, and `_quantile!`. For reason, see [this `StatsBase` comment](https://github.com/JuliaStats/StatsBase.jl/pull/586#issuecomment-653936354). For full discussion, see JuliaStats/StatsBase.jl#586 and [JuliaStats/Distributions.jl#1150](JuliaStats/Distributions.jl/issues/1150).

ron-wolf

Implement `pdf(d, missing)`

1

Be nice if `pdf(d, missing) = missing`. But we don't have that: ```julia julia> d=Distributions.Normal() Distributions.Normal{Float64}(μ=0.0, σ=1.0) julia> pdf(d, missing) ERROR: MethodError: no method matching pdf(::Distributions.Normal{Float64}, ::Missing) ```

ablaom

Hey there, While the `Product` Distribution framework is already quite useful, I feel there are several features missing that could be easily implemented. The first I can think about is...

gdalle