Open Access
February 2015 Testing the regularity of a smooth signal
Alexandra Carpentier
Bernoulli 21(1): 465-488 (February 2015). DOI: 10.3150/13-BEJ575

Abstract

We develop a test to determine whether a function lying in a fixed $L_{2}$-Sobolev-type ball of smoothness $t$, and generating a noisy signal, is in fact of a given smoothness $s\geq t$ or not. While it is impossible to construct a uniformly consistent test for this problem on every function of smoothness $t$, it becomes possible if we remove a sufficiently large region of the set of functions of smoothness $t$. The functions that we remove are functions of smoothness strictly smaller than $s$, but that are very close to $s$-smooth functions. A lower bound on the size of this region has been proved to be of order $n^{-t/(2t+1/2)}$, and in this paper, we provide a test that is consistent after the removal of a region of such a size. Even though the null hypothesis is composite, the size of the region we remove does not depend on the complexity of the null hypothesis.

Citation

Download Citation

Alexandra Carpentier. "Testing the regularity of a smooth signal." Bernoulli 21 (1) 465 - 488, February 2015. https://doi.org/10.3150/13-BEJ575

Information

Published: February 2015
First available in Project Euclid: 17 March 2015

zbMATH: 1320.94021
MathSciNet: MR3322327
Digital Object Identifier: 10.3150/13-BEJ575

Keywords: functional analysis , minimax bounds , non-parametric composite testing problem

Rights: Copyright © 2015 Bernoulli Society for Mathematical Statistics and Probability

Vol.21 • No. 1 • February 2015
Back to Top