A system for detecting monetization abuse by channels of content sharing networks is disclosed. The proposed system uses machine learning methods to generate a predictive model capable of being applied to channels to determine a likelihood that the channel is engaged in monetization abuse. Specifically, the proposed system may train a machine learning classifier using a plurality of training videos. The training may be based on generating clusters of videos and identifying which clusters are risky. The machine learning classifier may then be applied to existing and new channels to detect monetization abuse (e.g., channels uploading inappropriate content) and flag the potentially abusive channels for administrative review.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Chiarandini, Luca and Heldt, Lukasz, "MONETIZATION ABUSE DETECTION BASED ON CO-WATCH SIMILARITY", Technical Disclosure Commons, (April 27, 2020)