-
Notifications
You must be signed in to change notification settings - Fork 12
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Bandwidth selection in type_density and type_ridge #291
Conversation
Sorry, I forgot to check the density default checks more carefully hence the checks failed. Nevertheless I'll first wait for your feedback or changes before proceeding with updating the tests. |
This is fantastic. Thanks @zeileis! I agree with all of these changes. I'll quickly update the test snapshots on my side and push them so that the CI checks are passing. I can also think of one or two additions that we can add as part of this PR too. (E.g., I need to update the NEWS file.) I'll try to add those tonight. |
Excellent, thanks for wrapping all of this up! 👌 I cleaned up the corresponding branches now. I noticed that there is an occurrence of |
FWIW, I've pushed trivial things like this in the past and Grant has not gotten angry at me yet :) |
Thanks, fixed now! |
Def feel free to fix trivial issues directly (and apologies for the typo). |
Will do - no worries about the typo! |
Grant @grantmcdermott, this is a follow-up to #270 where I did the following in three separate commits:
joint.bw = "owm"
the default intype_density()
and enhanced the documentation. In particular, I added a new section explaining the practical considerations for choosing the differentjoint.bw
alternatives. I wrote this mostly so that we do not forget about it but feel free to shorten it (or remove it entirely) if you feel that this is too detailed.type_ridge()
documentation to the changes in thetype_density()
documentation.joint.bw = "mean"
instead of"owm"
. I find the latter too opaque and it over-emphasizes the weighting. I find"mean"
much more straightforward and the weighting is natural (to me) and well-explained in the documentation. If you feel strongly about not using"mean"
, feel free to revert this commit.If you keep the last commit, you way want to further adjust the tests. I have changed their code so that they run successfully but kept "owm" in the annotations so that the snapshots don't need updating, yet.
Closes #290