Seth Axen
Seth Axen
> Similarly, we could implement an `InverseWishartCholesky` to avoid Cholesky factorizing the input in `logpdf`. However, `rand` does not benefit in this case, since it would be implemented in terms...
One challenge is that `eltype` for a distribution `d` doesn't tell us what type `rand(d)` will return. The worst case here is `LKJCholesky`, for which the eltype is `Float64`, but...
> IMO `eltype` is a design flaw and should be removed completely. So I don't think it should block the PR. Without `eltype`, how would you recommend container eltype be...
> the less known `randtype` seems much clearer In which package is `randtype` defined? I have not heard of it. > It is intended to return the element type of...
I added a new API function `marginal`. While we have `component` for mixture distributions, we didn't have any API function for accessing marginals of a product distribution. And there are...
@devmotion this is ready for a review
This is really just waiting on a review. @devmotion?
@devmotion what do you think about this proposed docs restructure? > RE docs, currently `product_distribution` is documented in the multivariate distributions page, but with this PR, sometimes `product_distribution` will return...
Bump @devmotion
@devmotion I restructured the docs. I also noticed that, while `Product` is deprecated, its docstring was missing a deprecation warning, so I added it. I also figured the behavior of...