Abstract
The subject disclosure provides computer-implemented methods comprising training, using a holdout dataset, a false positive prediction model to generate a false positive score, the false positive score comprising a probability that an enforcement score generated by an enforcement model is a false positive; generating, from user action data of a first user, using the enforcement model, a first enforcement score; generating, from the user action data of a first user and the trained false positive prediction model, a first false positive score; and suppressing, using the first enforcement score and the first false positive score, an enforcement action on the first user.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
"IDENTIFICATION OF FALSE POSITIVES WITHIN SOCIAL MEDIA INTEGRITY ENFORCEMENT", Technical Disclosure Commons, (August 26, 2024)
https://www.tdcommons.org/dpubs_series/7308