Abstract
A recommendation system detects psychological reactance using explicit, implicit, and contextual behavioral signals and computes a per-user reactance score with temporal decay. A personalization intensity is determined per user and per content domain based on a domain-specific base intensity, the reactance score, and a context multiplier, subject to a minimum floor. The system modulates ranking in accordance with the intensity by adjusting feature-group weights and/or masking tiers of features. The system also generates an explanation whose framing corresponds to the computed intensity and truthfully reflects factors used for selection, including personal behavioral framing at higher intensities, neutral interest-based framing at mid intensities, and social proof or popularity framing at lower intensities. Per-user reactance state and histories may be stored and updated per request and per event, with periodic cleanup and monitoring of intensity distributions and engagement impact.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Anonymous, "Reactance-Aware Adaptive Personalization Intensity with Behavioral Detection and Dynamic Modulation in Recommendation Systems", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/10724