Selfnormalization for relevant inference with supremum-type statistics

2026-07-30

AI Summary

This study addresses the challenge of change-point testing in functional time series based on the supremum norm, a problem rendered difficult by the non-Hadamard differentiability of the norm, which invalidates conventional self-normalization techniques and yields limiting distributions dependent on the geometry of the extremal set and long-run covariance. To overcome these obstacles, this work proposes a novel framework that replaces the supremum norm with a smooth log-sum-exp approximation and integrates projection-based self-normalization with multiscale smoothing bias correction. The resulting procedure avoids the need to estimate nuisance parameters such as the extremal set geometry or long-run covariance, yielding a pivotal limiting distribution that depends solely on the change-point location. Under mild conditions, the method achieves asymptotically exact inference and is broadly applicable to a wide class of supremum-type statistics.

Abstract

We develop a selfnormalized approach to inference for relevant changes in functional time series measured by the supremum norm. The main difficulty is that the supremum norm is not Hadamard differentiable, so standard projection-based selfnormalization does not apply and the limiting distribution may depend on the geometry of the extremal set and the long-run covariance structure. We address this problem by replacing the supremum norm with a smooth log-sum-exp approximation and constructing a projected selfnormalizer from its derivative. The resulting statistic has an asymptotically pivotal distribution that is free of long-run covariance nuisance parameters and depends only on the break location. We derive explicit smoothing-bias expansions for both isolated nondegenerate extrema and extremal sets of positive measure. To avoid direct estimation of geometric quantities such as the number, curvature, or measure of the extrema, we combine several smoothing levels to cancel the leading bias terms. This yields an asymptotically exact test for relevant changes under mild regularity conditions. More generally, the proposed smoothing and bias-correction principles provide a framework for combining selfnormalization with supremum-type statistics in problems involving relevant hypotheses.

Problem

• self-normalization
• supremum-type statistics
• relevant inference
• functional time series
• Hadamard differentiability

Innovation

• self-normalization
• supremum-type statistics
• log-sum-exp approximation
• bias correction
• functional time series