Benjamin Mah
Benjamin Mah
Example diffs extracted: [Backout Data Collection Validation](https://docs.google.com/document/d/1akKmKFdGZgmyf-1h7nwx9EFGyNjPqljXAsrlLC--q74/edit?usp=sharing)
Sample dataset: https://drive.google.com/file/d/1LpUDZD3Ww8hnoylwrpPZqF8v5nxhrA_L/view?usp=sharing
[Updated Sample Dataset](https://github.com/user-attachments/files/16679512/inline_comment_dataset.json) @suhaibmujahid
[updated_dataset.json](https://github.com/user-attachments/files/16716364/inline_comment_dataset.json)
Sample dataset: [dataset.json](https://github.com/user-attachments/files/16751807/dataset.json)
Closing issue, as the metrics of the model with the features appear to be lower than the current model. See https://github.com/mozilla/bugbug/pull/4298#discussion_r2045398668.
> We should definitely include as "negative" examples in the Fenix component model training set, bugs that were put in Fenix by our bot but moved out of Fenix. Would...
After some investigation, there are two possible cases that we can consider: - originally a non-Fenix product -> `Fenix::` by BugBot -> a non-Fenix product by a Bugzilla user -...
Metrics of the model with the features included: [metrics.log](https://github.com/user-attachments/files/16086986/metrics_new.log) Consistently achieves 0.59 for accuracy and recall for all confidence thresholds.
> Instead of having specific features for this, you could make sure the function to restore the bug to its initial version also handles changes to product and component: Should...