No label defined (Q3747574)

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 3984407
Language Label Description Also known as
default for all languages
No label defined
    English
    No label defined
    scientific article; zbMATH DE number 3984407

      Statements

      0 references
      0 references
      1986
      0 references
      spectral density estimation
      0 references
      derivatives
      0 references
      kernel estimates
      0 references
      higher order kernels
      0 references
      choice of window widths
      0 references
      real-valued stationary process
      0 references
      spectral window
      0 references
      periodogram
      0 references
      integrated mean squared error
      0 references
      electroencephalography
      0 references
      neurophysiology
      0 references
      Let X(t), \(t=0,\pm 1,..\). be a real-valued stationary process with zero expectation and spectral density f(\(\lambda)\). The estimate proposed for \(f^{(\nu)}(\lambda)\), \(\lambda\in (0,\pi)\), is \[ f_{n,\nu}(\lambda)=\frac{1}{b^{\nu +1}}\cdot \sum^{n}_{j=1}I^{(2n)}(\frac{\pi j}{n})\cdot \int^{s_ j}_{s_{j-1}}W_{\nu}(\frac{\lambda -u}{b})du, \] where \(W_{\nu}\) is a spectral window, \(b=b(n)\) is a sequence of bandwidths, \(b\to 0\), nb\(\to \infty\), \(I^{(2n)}(\lambda)\) is the periodogram. Under some conditions on \(W_{\nu}\) it is proved that \[ E[f_{n,\nu}(\lambda)- f^{(\nu)}(\lambda)]^ 2= \] \[ c(k,\nu)n^{-2(k-\nu)/(2k+1)}[f^ 2(\lambda)\int^{1}_{-1}W^ 2_{\nu}(x)dx]^{2(k- \nu)/(2k+1)}\times \] \[ [\beta^ 2f^{(k)^ 2}(\lambda)]^{(2\nu +1)/(2k+1)}+o(n^{-2(k-\nu)/(2k+1)}), \] where c(k,\(\nu)\) is a constant depending only on k,\(\nu\), \(k\geq \nu +2\). The authors define a more general estimate, which is suitable for estimation in the interior as well as at the ends of [0,\(\pi\) ]. The main term of the asymptotical behaviour of the integrated mean squared error is obtained. Some applications to electroencephalography and neurophysiology are given.
      0 references

      Identifiers