Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

AUCPR Curve looks to be incorrect (due to incorrect recalls and precisions) #15637

Closed
wendycwong opened this issue Jul 14, 2023 · 0 comments
Closed
Assignees
Labels
Milestone

Comments

@wendycwong
Copy link
Collaborator

Support ticket: https://support.h2o.ai/helpdesk/tickets/105798.

Customer issue:

The .plot method for the h2o.model.metrics.binomial.H2OBinomialModelMetrics class seems to create an incorrect or misleading plot for type='pr'.
This looks to be due to the internal method _plot_pr assigning incorrect values to the variables recalls and precisions.
These variables are created on lines 574 and 575 in the binomial.py file on the Github repo, but I believe they actually reference the thresholds and recalls respectively.

image

image

@wendycwong wendycwong added the bug label Jul 14, 2023
@maurever maurever added this to the 3.42.0.2 milestone Jul 17, 2023
maurever added a commit that referenced this issue Jul 17, 2023
@maurever maurever changed the title AUCPR Curve looks to be incorrect (due to incorrect recalls and precisions) AUCPR Curve looks to be incorrect (due to incorrect recalls and precisions) [nocheck] Jul 20, 2023
@maurever maurever changed the title AUCPR Curve looks to be incorrect (due to incorrect recalls and precisions) [nocheck] AUCPR Curve looks to be incorrect (due to incorrect recalls and precisions) Jul 20, 2023
maurever added a commit that referenced this issue Jul 20, 2023
* GH-15637 fix aucpr plot bug

* regenerate image for doc
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

2 participants