Central limit theorem for weakly lacunary Walsh series (Q1078461): Difference between revisions

From MaRDI portal
Importer (talk | contribs)
Created a new Item
 
ReferenceBot (talk | contribs)
Changed an Item
(4 intermediate revisions by 3 users not shown)
Property / reviewed by
 
Property / reviewed by: Antónia Földes / rank
Normal rank
 
Property / reviewed by
 
Property / reviewed by: Antónia Földes / rank
 
Normal rank
Property / MaRDI profile type
 
Property / MaRDI profile type: MaRDI publication profile / rank
 
Normal rank
Property / cites work
 
Property / cites work: Central limit theorem for the Walsh system / rank
 
Normal rank
Property / cites work
 
Property / cites work: On Lacunary trigonometric series / rank
 
Normal rank
Property / cites work
 
Property / cites work: On Lacunary trigonometric series, II / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q4746556 / rank
 
Normal rank
links / mardi / namelinks / mardi / name
 

Revision as of 15:01, 17 June 2024

scientific article
Language Label Description Also known as
English
Central limit theorem for weakly lacunary Walsh series
scientific article

    Statements

    Central limit theorem for weakly lacunary Walsh series (English)
    0 references
    0 references
    0 references
    1985
    0 references
    Let \(x\in [0,1]\), \(r_ k(x)=sgn(\sin 2^ k\pi x)\), k-1,2,..., be the Rademacher functions. For each natural number n, \(n=\sum_{k}\epsilon_ k2^ k\) \((\epsilon_ k=0\) or \(\epsilon_ k=1)\), set \(\omega_ n(x)=\prod_{k}\{r_{k+1}(x)\}^{\epsilon_ k}.\) Moreover, let \(\omega_ 0(x)\equiv 1\). Then \(\{\omega_ n(x)\}\) is a Walsh system in Paley's ordering. In 1975 A. Földes proved the following central limit theorem: Theorem A. Let us be given a sequence of natural numbers \(\{n_ k\}\) and a sequence of non-negative numbers \(\{a_ k\}\), satisfying the conditions (1.1) \(n_{k+1}>n_ k(1+c\cdot k^{-\alpha})\), \(k=1,2,...\), (1.2) \(A^ 2_ N=\sum^{N}_{k=1}a^ 2_ k\to \infty\) as \(N\to \infty\), (1.3) \(a_ N=O(A_ N\cdot N^{-\alpha})\) as \(N\to \infty\), where c and \(\alpha\) are constants such that \(c>0\) and \(0\leq \alpha \leq\). Then for all t \[ (1.4)\quad \lim_{m\to \infty}| \{x:\quad x\in [0,1],\quad \sum^{N}_{k=1}a_ k\omega_{n_ k}(x)\leq t\cdot A_ N\}| =\frac{1}{\sqrt{2\pi}}\int^{t}\quad_{-\infty}\exp (- \frac{z^ 2}{2})dz, \] where the sign \(| \cdot |\) denotes the Lebesgue measure of a set. The aim of the present paper is to prove that estimate (1.3) is the best possible. More precisely the author proves that: Theorem B. For any fixed constants c and \(\alpha\) such that \(c>0\) and \(0\leq \alpha \leq\), there exists sequences \(\{n_ k\}\) and \(\{a_ k\}\), satisfying conditions (1.1), (1.2) and \((1.5)\quad a_ N=O(A_ N\cdot N^{-\alpha})\) as \(N\to \infty\), such that relation (1.4) is not satisfied.
    0 references
    0 references
    Rademacher functions
    0 references
    Walsh system
    0 references
    Paley's ordering
    0 references