Abstract
Online content providers typically provided personalized content recommendations. Content is available across many domains or verticals. Within each vertical, content can occur in clusters. A cluster is a set of content with correlated viewership. Certain metrics such as user engagement, user appeal, etc., apply to all content verticals while some metrics such as re-engagement, subscriptions, etc., apply only to specific verticals. The vertical-specific metrics are not comparable across verticals. This disclosure describes techniques to rank clusters within verticals that may have different and incomparable metrics or objectives. In a first pass, the clusters are ranked by metrics common across multiple verticals. In a second pass, the clusters with each vertical are re-ranked by metrics or objectives core to that vertical.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Gu, Zhiwei; Zhou, Will Xi; Liu, Jiahui; and Zhang, Zemin, "Universal Content Recommender", Technical Disclosure Commons, (June 10, 2022)
https://www.tdcommons.org/dpubs_series/5187