Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2602.19284 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917287797194752 |
|---|---|
| author | Wang, Yuhao Wang, Tengyao |
| author_facet | Wang, Yuhao Wang, Tengyao |
| contents | We propose a localized conformal model selection framework that integrates local adaptivity with post-selection validity for distribution-free prediction. By performing model selection symmetrically across calibration points using upper and lower surrogate intervals, we construct a data-dependent safe index set that contains the oracle model and preserves exchangeability. The resulting ensemble procedure retains exact finite-sample marginal coverage while adapting to spatial heterogeneity and model complexity. Simulations demonstrate substantial reductions in interval length compared to the best fixed model, especially in heterogeneous and low-noise settings. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_19284 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Localized conformal model selection Wang, Yuhao Wang, Tengyao Methodology 62G15 We propose a localized conformal model selection framework that integrates local adaptivity with post-selection validity for distribution-free prediction. By performing model selection symmetrically across calibration points using upper and lower surrogate intervals, we construct a data-dependent safe index set that contains the oracle model and preserves exchangeability. The resulting ensemble procedure retains exact finite-sample marginal coverage while adapting to spatial heterogeneity and model complexity. Simulations demonstrate substantial reductions in interval length compared to the best fixed model, especially in heterogeneous and low-noise settings. |
| title | Localized conformal model selection |
| topic | Methodology 62G15 |
| url | https://arxiv.org/abs/2602.19284 |