Skip to content
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

Matching package dependency #41

Closed
MartinBeal opened this issue Apr 12, 2022 · 6 comments
Closed

Matching package dependency #41

MartinBeal opened this issue Apr 12, 2022 · 6 comments
Assignees

Comments

@MartinBeal
Copy link
Collaborator

MartinBeal commented Apr 12, 2022

As you can see here, the Matching package is not currently available on CRAN. Because the track2KBA::indEffectTest() function uses Matching::ks.boot(), our package is at risk of being archived.

@VirginiaMorera can you think of an alternative method for running this bootstrap that wouldn't require the Matching package?

We have until April 23rd to fix this and pass the CRAN tests or our package will be removed from CRAN.

@MartinBeal MartinBeal self-assigned this Apr 12, 2022
@steffenoppel
Copy link
Collaborator

steffenoppel commented Apr 12, 2022

Could we simply replace it with 'ks.test' or provide a bootstrapped ks-test that does not require a deprecated package, e.g. adopting the code from this thread or from this website? Might not be very elegant - but better than being archived!

@MartinBeal
Copy link
Collaborator Author

MartinBeal commented Apr 20, 2022

'Matching' has become available again, so I am waiting to hear back from CRAN maintainers on whether or not we need to make any further changes. Worst case, I've made a branch 'no_match_depend' which replaces the bootstrapped ks.test with a simple ks.test which can be used in case we need it.

Will let you know what CRAN says.

@VirginiaMorera
Copy link
Contributor

Sorry I haven't been able to look into this until now. The reason I used the matching package ks test is that it does not assume independence of samples (since the kernels involved in the within individual overlaps will also be involved in the between individual overlaps) and also does not assume continuous distributions (since the overlap values are percentages). I seem to remember there was not a big difference between the bootstrapped KS test and the normal KS test, but it seemed more appropriate. Let's see if matching can still be a solution, if not, I'd at least issue a warning with the result if we use the non bootstrapped ks test? I'll try to think of an alternative way of doing this

@MartinBeal
Copy link
Collaborator Author

Ok so still not 100% certain (we'll see come Friday) but I BELIEVE the package is not currently at risk of archiving, since the Matching package is again available on CRAN.

@MartinBeal
Copy link
Collaborator Author

OK so just FYI our package has been taken off CRAN until we pass the checks. I re-submitted it last friday but it came back with a tiny note so I have fixed that and re-submitted today. Hopefully should be back online late this week.

@MartinBeal
Copy link
Collaborator Author

Ok, we're back online. For now no need to worry about removing the Matching dependency, as that package appears to also be actively maintained.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants