Surrogate-assisted optimal sampling for risk prediction under measurement constraints
Sunhyun Park, Seong-ho Lee · 2026
Read Surrogate-assisted optimal sampling for risk prediction under measurement constraints by Sunhyun Park, Seong-ho Lee (2026) 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-2606-03477.txt.
Description
Park and Lee design a smarter way to choose which cases get costly outcome checks. In many risk studies, cheap data cover a large target group. A surrogate outcome is also known for that whole group. The true outcome costs much more to measure, so only a subset can receive it. The sampling choice shapes how good the final risk model can be.
The surrogate flags confirmed positive cases on its own. For surrogate-negative cases, the true outcome stays unknown unless measured. The framework picks an optimal design that cuts expected cross-entropy loss. That design feeds a weighted estimator for the risk model. It needs only covariates, the surrogate, and a first rough estimator. The authors prove consistency and asymptotic normality for the estimator. Simulations and two real datasets show better prediction than rival designs. The edge holds even when the surrogate is flawed.
Catalog record
| CREATOR | Sunhyun Park, Seong-ho Lee |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 67 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.03477