9
$\begingroup$

I recently came across the following question on Stackoverflow (after getting stuck on the same issue independently):

https://stackoverflow.com/questions/60439187/why-grpreg-library-and-gglasso-library-in-r-are-giving-different-results-for-gro

It turns out that the answer is nuanced and relates to the impact of group-wise orthogonalization on the group lasso penalty.

I think it is worth writing down the answer to the above question; however I don't want to answer on Stackoverflow, because math mode seems not to be enabled and also it seems that the question would be of a lot more interest to the cross-validated community.

Given the above, how should I proceed?

  1. Open a similar (better phrased) question on cross-validated and answer myself and then link to the response as a comment on the Stackoverflow post?
  2. Bite the bullet and reply directly on Stackoverflow?
  3. Is there any other reasonable option?
$\endgroup$
2
  • 11
    $\begingroup$ If the answer is about the statistical principles at issue & how they play out in that situation, then the Q belongs here, not on SO. Use option #1, & vote to close there (it won't be possible to migrate it at this point). If the answer is about how the function works / how to use the code, then it belongs on SO, not here. Use option #2. $\endgroup$ – gung - Reinstate Monica Oct 19 '20 at 11:35
  • 1
    $\begingroup$ It's definitely option #1, thanks! $\endgroup$ – air Oct 19 '20 at 19:07
8
$\begingroup$

Questions and answers are under CC BY-SA 4.0

That means that you can literally copy the question over if you make it explicitly clear in the text its attribution (a link to the original question and user, and probably quoting the question is more than acceptable).

You can also paraphrase the question (linking the original as the inspiration would be fine) in better terms.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .