No label defined (Q3347138)

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 3881719
Language Label Description Also known as
default for all languages
No label defined
    English
    No label defined
    scientific article; zbMATH DE number 3881719

      Statements

      1984
      0 references
      asymptotic normality
      0 references
      regression function
      0 references
      stopping variable
      0 references
      Let \(\{(X_ i,Y_ i)\), \(i\geq 1\}\) be independent copies of a two dimensional random variable (X,Y) with joint density f, and let \(m(x)=E(Y| X=x)\) be the regression function. Let N(t) be a random number of observations in time (0,t] (possibly a stopping variable). The author considers the estimator of the regression function m(x) defined by \[ m_{N(t)}(x)=[\sum^{N(t)}_{i=1}Y_ ih_ i^{-1}K((x-X_ i)h_ i^{-1})]/\sum^{N(t)}_{i=1}h_ i^{-1}K(x-X_ i)h_ i^{-1}), \] where K(u) is a density function on (-\(\infty,\infty)\) with suitable growth conditions and \(\{h_ n\}\) is a monotonically decreasing sequence of positive numbers. For distinct values \(x_ 1,x_ 2,...,x_ k\) the author proves that under certain regularity conditions, \[ \sqrt{N(t)h_{N(t)}}(m_{N(t)}(x_ 1)-m(x_ 1),...,m_{N(t)}(x_ k)-m(x_ k)) \] converges in distribution to a normal law with zero mean vector and suitable dispersion matrix.
      0 references
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references