Some mixing properties of tukey's 3R smoother<sup>†</sup> (Q3314653)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 3844733
Language Label Description Also known as
default for all languages
No label defined
    English
    Some mixing properties of tukey's 3R smoother<sup>†</sup>
    scientific article; zbMATH DE number 3844733

      Statements

      Some mixing properties of tukey's 3R smoother<sup>†</sup> (English)
      0 references
      0 references
      1984
      0 references
      median
      0 references
      strong mixing
      0 references
      mixing rate
      0 references
      0 references
      \textit{J. W. Tukey's} 3R (''running median'') smoothing algorithm [Exploratory data analysis. (1977; Zbl 0409.62003)] is defined as follows: Given a sequence \(X:=(X_ k,k\in {\mathbb{Z}})\) of real-valued random variables, define the random variables \(X_ k^{(n)}\), \(k\in {\mathbb{Z}}\), \(n=0,1,2,..\). by \(X_ k^{(0)}:=X_ k \forall k\in {\mathbb{Z}}\) and \(X_ k^{(n)}:=med(X_{k-1}\!\!^{(n-1)},X_ k^{(n- 1)},X_{k+1}\!\!^{(n-1)})\), \(\forall k\in {\mathbb{Z}}\), \(\forall n\geq 1\), where ''med'' denotes the median; and define the sequence \(X^{(\infty)}:=(X_ k^{(\infty)},k\in {\mathbb{Z}})\) of random variables by \(X_ k^{(\infty)}:=\lim_{n\to \infty}X_ k^{(n)}\), \(\forall k\in {\mathbb{Z}}\) (if this limit exists a.s.). The sequence \(X^{(\infty)}\) is by definition the output of the 3R smoother if X is the input. If X is strictly stationary and satisfies the (Rosenblatt) strong mixing condition, then the same is true for \(X^{(\infty)}\). If the ''mixing rate'' for X is \(0(n^{-\theta})\) for some \(\theta>0\), then the mixing rate for \(X^{(\infty)}\) is \(0(n^{-\theta +\epsilon}) \forall \epsilon>0\). If X is i.i.d. then the mixing rate for \(X^{(\infty)}\) is exponential. If X is i.i.d. with a continuous (that is, non-atomic) marginal distribution, then \(X^{(\infty)}\) fails to be \(\phi\)-mixing. Thus two questions of \textit{C. L. Mallows} [Smoothing techniques for curve estimation. Lect. Notes Math. 757, 77-90 (1979; Zbl 0427.93050)] are answered.
      0 references

      Identifiers