Properties of local orthonormal systems, Part II: Geometric characterization of Bernstein inequalities

    Research output: Contribution to journalArticlepeer-review

    Abstract

    Let (Ω,ℱ,P) be a probability space and let (ℱn)n=0 be a binary filtration. i.e. exactly one atom of ℱn−1 is divided into two atoms of ℱn without any restriction on their respective measures. Additionally, denote the collection of atoms corresponding to this filtration by A. Let S⊂L(Ω) be a finite-dimensional linear subspace, having an additional stability property on atoms A. For these data, we consider two dictionaries: • C={f⋅1A:f∈S,A∈A}, • Φ – a local orthonormal system generated by S and the filtration (ℱn)n=0. Let Lp(S)=span¯Lp(Ω)C=span¯Lp(Ω)Φ, with 1<p<∞. We are interested in approximation spaces Aqα(Lp(S),C) and Aqα(Lp(S),Φ), corresponding to the best n-term approximation in Lp(S) by elements of C and Φ, respectively, where α>0 and 0<q≤∞. It is known that in the classical Haar case, i.e. when S=span(1[0,1]) and the binary filtration (ℱn)n=0 is dyadic (that is, an atom A∈A is divided into two new atoms of equal measure), we have Aqα(Lp(S),Φ)=Aqα(Lp(S),C), cf. P. Petrushev (2003). This motivates us to ask the question whether this equality is true in the general setting described above. The answer to this question is governed by the validity of a specific Bernstein type inequality BI(A,S,p,τ), with parameters 1<p<∞, 0<τ<p. The main result of this paper is a geometric characterization of this type of Bernstein inequality BI(A,S,p,τ), i.e. a characterization in terms of the behavior of functions from the space S on atoms A and rings ℛ={A∖B:A,B∈A,B⊂A}∖A. We specialize this general result to some examples of interest, including general Haar systems and spaces S consisting of (multivariate) polynomials.

    Original languageEnglish
    Article number106149
    Number of pages40
    JournalJournal of Approximation Theory
    Volume308
    DOIs
    Publication statusPublished - Jun 2025

    Fields of science

    • 101002 Analysis
    • 101032 Functional analysis

    JKU Focus areas

    • Digital Transformation

    Cite this