A Bayesian Model for Multi-stage Censoring Thanks: Co-last
Shuvom Sadhuka, Sophia Lin, Bonnie Berger, Emma Pierson · 2025
Read A Bayesian Model for Multi-stage Censoring Thanks: Co-last by Shuvom Sadhuka, Sophia Lin, Bonnie Berger, Emma Pierson (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-11684.txt.
Description
Shuvom Sadhuka, Sophia Lin, Bonnie Berger, and Emma Pierson model a trap in medical data. Patients pass through funnels of tests. The final truth is known only for the few who reach the end. A breast exam leads to a mammogram for some. A biopsy follows for fewer still, and only it reveals the outcome.
Because outcomes are missing by design, risk estimates get biased. The bias hits underserved groups hardest, since their results are censored more often. A Bayesian model built for funnels fixes much of this. In synthetic tests it recovers the true values. It also predicts outcomes for censored patients better than baselines. Fairer risk estimates in staged screening are the prize.
Catalog record
| CREATOR | Shuvom Sadhuka, Sophia Lin, Bonnie Berger, Emma Pierson |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 46 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-11684