Benjamin Mah

> > > 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...

With the feature (highest threshold): pre rec spe f1 geo iba sup avg / total 0.57 0.50 1.00 0.53 0.54 0.49 8986 Without the feature (highest threshold): pre rec spe...

> With the feature (highest threshold): pre rec spe f1 geo iba sup avg / total 0.57 0.50 1.00 0.53 0.54 0.49 8986 > > Without the feature (highest threshold):...

Still a WIP. Metrics collected from SMOTE can be found here: [metrics.log](https://github.com/user-attachments/files/16029517/metrics.log) Takeaways: - SMOTE increases the training time to around 2 hours (as opposed to the current 30-40 minute...

@suhaibmujahid Since this PR is mainly experimental, and the results were proven to be much worse, I think we can close this. WDYT?

Hello, I'd like to work on this issue, if possible!

Metrics of the newly trained model: [metrics.log](https://github.com/user-attachments/files/15919477/metrics.log)

> Do you see significant improvement when adding this feature? I've previously attached the metrics of the model here: > Metrics of the newly trained model: [metrics.log](https://github.com/user-attachments/files/15919477/metrics.log) Here are the...

I've converted this PR to a draft, as I realized there still needs some polishing to do with the extraction of file paths. For example, there are cases where it...

Current metrics: [metrics.log](https://github.com/user-attachments/files/16316806/metrics.log) Seems to perform slightly worse than the current model and `python3 -m scripts.bug_classifier component --bug-id 1902245` classifies this bug as `Core::Widget: Gtk` (which is incorrect). It is...