Timeline for Models have high accuracy but very low AUC PR curve
Current License: CC BY-SA 4.0
3 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 17, 2020 at 21:43 | comment | added | bolleke | Thank for the link to the answer, it helps, although I also get a very low result for the AUC ROC curve. Which is weird because when I use pROC's auc() function on my model, results are higher. So I know I'm doing something wrong, I just don't know where | |
| May 17, 2020 at 21:23 | comment | added | EdM | Note that the shape of a precision-recall curve can be more informative than the area under it. See this page and pages linked from there. Also, interpreting the AUPRC value isn't quite so straightforward as is interpreting the AUROC value. It might help to show both the PR curve and the corresponding ROC curve to provide more context. | |
| May 17, 2020 at 20:53 | history | asked | bolleke | CC BY-SA 4.0 |