Large-scale linear hypothesis testing for high-dimensional online M-estimation
Abstract
With the growing prevalence of streaming data, renewable estimation has become an important tool for updating statistical analyses without retaining historical raw observations. We study estimation and linear hypothesis testing in high-dimensional online M-estimation. Existing renewable procedures based on quadratic approximations can incur first-order approximation errors because the gradient of the unpenalized loss generally does not vanish at a preceding penalized estimator, posing a particular challenge for valid inference. To address this issue, we propose a corrected renewable loss that preserves historical first-order information and combine it with partial folded concave regularization and an iterative local linear approximation. We establish finite sample ℓ1- and ℓ2-error bounds, together with contraction bounds and strong oracle properties for both the constrained and unconstrained iterative estimators. Building on these results, we derive oracle Bahadur representations and establish asymptotically valid Wald and score tests for general linear hypotheses, allowing both the testing-oracle dimension and the number of restrictions to diverge. We further characterize their asymptotic power under local alternatives. The proposed procedure requires only recursively updated summaries while recovering the first-order inferential behavior of its pooled-data oracle counterpart. Simulation studies and real-data applications demonstrate the finite-sample performance of the proposed estimation and testing procedures.
Related
- Design-Driven Inference for Online Experiments: Jointly Optimal Evidence Collection, Testing, and Estimation
- Nonparametric Hypothesis Testing of High-dimensional Clustering With Application to Single-cell RNA Data
- Optimal Linear Functional Estimation for Proportionally High-Dimensional Linear Models
- A Model of Multiple Hypothesis Testing
- Principal-Agent Hypothesis Testing
BibTeX
@misc{ham2026largescale,
title = {{Large-scale linear hypothesis testing for high-dimensional online M-estimation}},
author = {Daeyoung Ham and Myeonghun Yu and Tate Jacobson},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.07418},
eprint = {2610.07418},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.07418},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Large-scale linear hypothesis testing for high-dimensional online M-estimation
AU - Daeyoung Ham
AU - Myeonghun Yu
AU - Tate Jacobson
PY - 2026
DA - 2026/10/05
PB - arXiv
DO - 10.48550/arxiv.2610.07418
UR - https://arxiv.org/abs/2610.07418
AB - With the growing prevalence of streaming data, renewable estimation has become an important tool for updating statistical analyses without retaining historical raw observations. We study estimation and linear hypothesis testing in high-dimensional online M-estimation. Existing renewable procedures based on quadratic approximations can incur first-order approximation errors because the gradient of the unpenalized loss generally does not vanish at a preceding penalized estimator, posing a particular challenge for valid inference. To address this issue, we propose a corrected renewable loss that preserves historical first-order information and combine it with partial folded concave regularization and an iterative local linear approximation. We establish finite sample ℓ_1- and ℓ_2-error bounds, together with contraction bounds and strong oracle properties for both the constrained and unconstrained iterative estimators. Building on these results, we derive oracle Bahadur representations and establish asymptotically valid Wald and score tests for general linear hypotheses, allowing both the testing-oracle dimension and the number of restrictions to diverge. We further characterize their asymptotic power under local alternatives. The proposed procedure requires only recursively updated summaries while recovering the first-order inferential behavior of its pooled-data oracle counterpart. Simulation studies and real-data applications demonstrate the finite-sample performance of the proposed estimation and testing procedures.
ER -