Update Predicted Change in F-Measure

[zakih2]

Pull Request

• I confirm that I have read the CONTRIBUTING.md document.
• I confirm that this contribution is made under an MIT license and that I have the authority necessary to make this contribution on behalf of its copyright owner.
• I confirm that I did not use any AI-based coding assistance (such as, but not limited to, GitHub Copilot).

Description

Updated PCF (base.py) to work as specified in the published implementation.

[chschroeder]

Thanks for the PR!

As discussed, this corrects a deviation of the current implementation from the original paper.

A short summary of our e-mail correspondence so that others can follow:

• Humza Zaki, Vijay Manchiraju, and Dr. Michael Bloodgood found an issue in the implementation the of Predicted Change in F-Measure stopping criterion, where it deviates from the logic of the original paper. Contrary to the paper, the implementation also considered agreement over the negative class.
• Since Dr. Bloodgood one of the authors of this strategy, nobody else could give a more accurate assessment here.
• They investigated the behavior of the observed implementation compared to the original implementation and noted a singificant difference.
• Moreover, they already proposed a fix. Very appreciated :).
• Since they already invested considerable time in finding and evaluating the issue, I offered to do the documentation work (docstrings, changelog, and docs).

I will add the documentation parts and then accept the PR.

[chschroeder]

Done. Thanks again for your efforts in investigating this 👍.

[zakih2]

No problem. Thank you for considering our suggestions. *- Humza Zaki* The College of New Jersey - Class of 2024 Computer Science Major ***@***.***

…

On Wed, Dec 20, 2023 at 4:12 PM Christopher Schröder < ***@***.***> wrote: Done. Thanks again for your efforts in investigating this 👍. — Reply to this email directly, view it on GitHub <#51 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AYNGO46L2HRPYUA373UYRCLYKNIFBAVCNFSM6AAAAABA3ZAL56VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNRVGE2DKMZZGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>