HB-PVI: A Hierarchical Bayesian Personalization and Value-of-Information Framework for Complex Activity Recognition
Read the original on arXiv Machine Learning →HB‑PVI is a hierarchical Bayesian framework that jointly models participant heterogeneity, the benefits and harms of four personalization mechanisms, and the economic value of acquiring additional labels for complex activity recognition. In a 47‑participant MUSIC‑CAR cohort, the framework used a sequential‑Monte‑Carlo updater, a Student‑t hierarchical gain model, and a one‑step expected‑value‑of‑sample‑information stopping rule. The results showed that, under realistic cost and benefit thresholds, the policy avoided labeling entirely, matching an always‑stop strategy while maintaining near‑optimal predictive performance and reducing labeling effort by 100%.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.