One-sided CIs for AUCs

[Re2124]

Is your feature request related to a problem? Please describe.
I would like to calculate one-sided CIs for AUC values in order to test one-tailed hypothesis.
(This is quite common in by research area. Many researchers would benefit from this added feature)

Describe the solution you'd like
So far, I found no possibility to do so using the pROC package.

Describe alternatives you've considered
It would be great if someone could implement a solution, where one could choose between one-sided (greater, less) and two-sided.
(Unfortunately, I do not have the knowledge to try to do ist myself)

Additional context
/

[xrobin]

Interesting, I never thought of that.

Can you point to any litterature that shows how one sided CIs are computed in the case of ROC curves?

I'd certainly welcome PR requests for this feature.

[Re2124]

Unfortunately no. It seams, however, be a problem also other researchers face: https://stats.stackexchange.com/questions/476499/one-sided-test-for-auc-area-under-the-curve-for-roc

This author wrote at least something about one-sided testing for AUC: https://dpc10ster.github.io/RJafrocBook/hypothesis-testing.html#hypothesis-testing-how-much-finished