Carlos Parada
Carlos Parada
> No, that's not correct. Usually it's MLE but it does not have to be (stated also in the docs: https://juliastats.org/Distributions.jl/latest/fit/). Sorry, I meant it should default to `fit_mle` unless...
> I am not really sure if the joint MLE of location and scale of the Cauchy distribution is biased. From my numerical experiments it looks like the bias of...
> Test seem to be complete and pass. So the final (main) question is: Do we actually want to add more constructors? Maybe it would be better to switch to...
I think we definitely need this; I'm trying to write up a package for queuing models and this is messing me up right now :sweat_smile: (Customers often have a wait...
@oschulz @joshday still interested in this? Would make plotting a lot easier with ASH.
Oh, that's why! I could've sworn I'd made a PR for this at some point. Can you reopen it? I don't have permission.
> Hi @joshday can you please describe how you were planning to implement this? I believe @itsdebartha would be interested in doing this!
We could use git subtrees or git submodules for a built-in system to handle this.
I'm actually not sure if separating this into separate packages is even necessary post-1.9, given the big improvements in compile times. Especially given there's a lot more on the way...
> Maybe one should just use a special set of weighted observations, i.e., `condition(model, observations)` where the weights are included in `observations`, to keep the API consistent. > > Related:...