Abstract
Techniques are described for recommending creator content using parasocial relationship lifecycle modeling. Interaction data are used to compute signals for each user-creator pair including cross-topic engagement consistency, engagement differential versus other creators within the same topics, and catalog coverage. A pair may be classified as parasocial based on thresholding these signals. A state machine assigns a lifecycle phase selected from discovery, rising attachment, peak attachment, fatigue, recovery, and disengagement using engagement level and engagement trends over time. Per-request ranking applies phase-specific strategies and exposure limits, selects a diversity dimension between topic diversity and creator diversity based on whether the user has one or multiple peak creators, and applies proactive fatigue prevention via exposure multipliers triggered by early warning decline conditions. Batch processing updates stored relationship state for use during recommendation requests.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Anonymous, "Parasocial Relationship Lifecycle Modeling for Creator Recommendations with Phase-Specific Ranking Strategies", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/10725