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

Fix #330 #344

Merged
merged 9 commits into from
Jul 27, 2022
Merged

Fix #330 #344

merged 9 commits into from
Jul 27, 2022

Conversation

fweber144
Copy link
Collaborator

This fixes issue #330 completely (commit 23e7101 already fixed it partly, i.e., as far as possible at that time).

Note that in contrast to what I assumed in the TODO comment added by 23e7101, the auc() function actually seems to have expected the CV fold weights in the third column of its input matrix. The user-supplied observation weights were simply ignored for the "auc" statistic before 23e7101.

efficiency (this is an equivalent transformation because of `weights / sum(weights)` which is performed directly afterwards).
… `"auc"` to a common one. This

is only a refactor (the lines `n_notna <- sum(d_test$weights)` and `weights <- n_notna * weights / sum(weights)` do not harm in case of the `"auc"` because the constant scalar factor `n_notna / sum(weights)` cancels out in `auc()`).
correct, but for consistency, it would have to be added to the `"acc"` (= `"pctcorr"`) case, too. And that would clutter the help page. But most importantly, it's quite obvious how AUC and accuracy are calculated in case of the "aggregated" binomial family.
correspondingly, element `<vsel_object>$summaries$<sub[[k]]_or_ref>$w` to `<vsel_object>$summaries$<sub[[k]]_or_ref>$wcv`) to avoid confusion in the future.
of the function definition. This seems to be an R convention.

Enhance another inline comment.
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

Successfully merging this pull request may close these issues.

None yet

1 participant