arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.19284stat.ME
收藏

局部共形模型选择

Localized conformal model selection

Tsinghua University(清华大学);Shanghai Qizhi Institute(上海启智研究院);London School of Economics and Political Science(伦敦政治经济学院)

Yuhao Wang, Tengyao Wang

AI总结:

本文提出局部共形模型选择方法,通过结合局部适应性和后选择有效性,实现无分布预测,并在异质性和低噪声环境下显著减少预测区间长度。

摘要:

我们提出了一种局部共形模型选择框架,该框架将局部适应性与后选择有效性相结合,用于无分布预测。通过在校准点上对称地使用上界和下界替代区间进行模型选择,我们构建了一个数据依赖的安全索引集,该集合包含oracle模型并保持可交换性。所得到的集合过程在保留精确有限样本边际覆盖的同时,能够适应空间异质性和模型复杂性。模拟结果表明,与最佳固定模型相比,区间长度显著减少,尤其是在异质性和低噪声设置中。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑