-
Notifications
You must be signed in to change notification settings - Fork 311
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
Evaluate multiple output component model #906
base: master
Are you sure you want to change the base?
Conversation
Codecov Report
@@ Coverage Diff @@
## master #906 +/- ##
==========================================
- Coverage 53.03% 52.85% -0.18%
==========================================
Files 81 81
Lines 5461 5487 +26
==========================================
+ Hits 2896 2900 +4
- Misses 2565 2587 +22
Continue to review full report at Codecov.
|
No confidence threshold - 12463 classified for product model
No confidence threshold - 12463 classified for component model
@marco-c I have a trained model saved in case you need to see full results. |
These lines should not be there, as these components should be ignored:
|
The results seem worse than the current model (can you post the results of the current model?), so perhaps this is not a good idea. The classifier chain might work better in this case. |
@marco-c we cannot use classifier chain here , as it is used to combine binary classifiers which is not our case i think as written here https://scikit-learn.org/stable/modules/multiclass.html#classifier-chain |
@marco-c The results of the current model for no threshold case..
|
I think these aren't ignored because these inverse mapping contains both General and Untriaged. bugbug/bugbug/models/component.py Lines 51 to 54 in 8b59321
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Will need to be rebased but I will keep this open in case we want to pick it up again.
#224
We are calculating metrics for each output(i.e., product, component and conflated_component) separately.