# Is there any intuition for the PRAUC metric?

**URL:** <https://discourse.datamethods.org/t/is-there-any-intuition-for-the-prauc-metric/4237>\
**Category:** accuracy\
**Tags:** interpretation\
**Created:** [March 24, 2021, 11:48am UTC](https://discourse.datamethods.org/t/is-there-any-intuition-for-the-prauc-metric/4237 "2021-03-24T11:48:47Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Uriah](https://discourse.datamethods.org/user_avatar/discourse.datamethods.org/uriah/32/1280_2.png) [@Uriah](https://discourse.datamethods.org/u/Uriah)\
**Post date:** [March 24, 2021, 11:48am UTC](https://discourse.datamethods.org/t/is-there-any-intuition-for-the-prauc-metric/4237/1 "2021-03-24T11:48:47Z")

</div>

Hello everyone, this is my first post here.

I’m familiar with the intuition behind the ROCAUC thanks to the c-index and I find it usefull in order to understand how well my model can discriminate between events and non-events.

I heared some reccomendations about the PRAUC metric, which is basically the area under the Precision-Recall curve. I can understand why higher PRAUC is better, but that’s about it.

I wonder if there’s any intuitive example for what the PRAUC metric really stands for.