Say that I have a set of training examples with binary 0-1 labels, and corresponding regression labels ranging from 0 to 1. Normally, to compute AUC I've used the ROCR package in R, which also allows me to compute the optimal cutoff point. In addition to these labels and predictions, I also have a weight column which ranges from 0 to 1 and signifies the importance of each example. I was wondering how I could compute the AUC and optimal cutoff point of these examples, taking these weights into account.


1 Answer 1


Not sure if this question is still valid, but you can use PRROC package in R for weighted AUC computations.

  • 1
    $\begingroup$ glmnet package also has it $\endgroup$
    – seanv507
    Dec 12, 2018 at 18:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.