Quantifying uncertainty and stability among highly correlated predictors: a subspace perspective

Xiaozhu Zhang, Department of Statistics, University of Washington · 2025

Read Quantifying uncertainty and stability among highly correlated predictors: a subspace perspective by Xiaozhu Zhang, Department of Statistics, University of Washington (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-2505-06760.txt.

Introduction

Xiaozhu Zhang and co-authors rethink feature choice when predictors move together in a 2025 paper on statistics. In many data sets, features are near copies of each other. Counting shared picks then misleads: two models can share no features yet predict alike.

Their fix is to judge models by the space their chosen features span. Near spaces count as near models, a smooth measure that fits tangled data. Stability and false alarms get the same smooth treatment. The team builds a stable selection method on this base and proves its behavior. When features are not tangled, the new measures match the old counting ones. The frame suits genomics and finance, where tangles are the norm.

Catalog record

CREATORXiaozhu Zhang, Department of Statistics, University of Washington
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE239 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-06760

Browse the full library · About the Tower of Babel