Inference without compatibility: Using exponential weighting for inference on a parameter of a linear model

Michael Law, Yaacov Ritov

Research output: Contribution to journalArticlepeer-review

3 Scopus citations

Abstract

consider hypotheses testing problems for three parameters in high-dimensional linear models with minimal sparsity assumptions of their type but without any compatibility conditions. Under this framework, we construct the first vn-consistent estimators for low-dimensional coefficients, the signal strength, and the noise level. We support our results using numerical simulations and provide comparisons with other estimators.

Original languageEnglish
Pages (from-to)1467-1495
Number of pages29
JournalBernoulli
Volume27
Issue number3
DOIs
StatePublished - Aug 2021
Externally publishedYes

Bibliographical note

Publisher Copyright:
© 2021 International Statistical Institute. All rights reserved.

Keywords

  • Compatibility Condition
  • Exponential Weighting
  • Inference
  • Lasso

Fingerprint

Dive into the research topics of 'Inference without compatibility: Using exponential weighting for inference on a parameter of a linear model'. Together they form a unique fingerprint.

Cite this