0
$\begingroup$

On the test set of a binary classification problem, the p25, p50 and p75 of the predictions are very close to each other (e.g. 0.123).

Is it possible that my model can achieve a high AUC-ROC (e.g. 0.85) despite giving the same probability prediction for almost the rows?

The data is imbalanced.

$\endgroup$
3
  • $\begingroup$ If it is a imbalanced class it can happen $\endgroup$ Commented Apr 6, 2022 at 18:37
  • $\begingroup$ "very close to each other", or "same for nearly all rows"? Some more details on the distribution of predicted probabilities, please. Also, what kind of model? $\endgroup$
    – Ben Reiniger
    Commented Apr 7, 2022 at 0:01
  • 1
    $\begingroup$ ROC only cares about ranking. So 0.1231 and 0.1232 are as far apart as -100 and +1000: 1 rank. Can you show the actual values? $\endgroup$
    – Calimo
    Commented Apr 7, 2022 at 5:38

0

Your Answer

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

Browse other questions tagged or ask your own question.