Mitchell O'Hara-Wild

> I'm currently doing a project where an `ARFIMA()` function would be very useful. Any generalisations that should be made from `forecast::arfima()`?

Are you able to provide the `data` object so that I can reproduce this issue? The issue may be related to having weekly data.

A `fable.tscount` would be great and I'm happy to help you write this if it is useful for you.

@robjhyndman This is coming from here (which comes from `forecast::auto.arima()`) https://github.com/tidyverts/fable/blob/64b36c6df3285e33c912991a1cc885604bbb3158/R/arima.R#L153 Any reason why `npar` is the number of estimated coefficients `+1`?

Also @TimothyHyndman, when you say "Refit ARIMA with two more observations", you're refitting the model with only two observations (not two more observations). You'll need to refit the model with...

Got it, thanks. https://github.com/tidyverts/fable/blob/64b36c6df3285e33c912991a1cc885604bbb3158/R/arima.R#L159 Then in this case (`n=2`, `d=0`, `D=0`), `nstar = 2` giving division by `(nstar - npar - 1) = 0`. Is this correct?

At this stage there is no functionality for calculating impulse responses for VAR models, but fundamentally there is nothing preventing that functionality from being added. The complexity with adding this...

The implementation of VAR in this package does not use the implementation from the `vars` package, so you won't be able to use `vars::irf()` on the mable to get the...

This seems to occur for all leap years, and is likely the purpose of this commented line: https://github.com/tidyverse/lubridate/blob/6f26b02de432cd9373ad4ce7766c36eacfc29918/src/update.cpp#L304 Looks like it was removed when resolving issue #567. The leap year...

Interested to see how this goes, please feel free to ask any questions for integration with fable. For full integration with fable you could also consider a wrapper package (much...