Estimation of the location and exponent of the spectral singularity of a long memory process (Q4828170)

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 2118732
Language Label Description Also known as
default for all languages
No label defined
    English
    Estimation of the location and exponent of the spectral singularity of a long memory process
    scientific article; zbMATH DE number 2118732

      Statements

      Estimation of the location and exponent of the spectral singularity of a long memory process (English)
      0 references
      0 references
      0 references
      24 November 2004
      0 references
      long memory
      0 references
      fractionally differenced processes
      0 references
      semiparametric estimation
      0 references
      The authors consider a covariance stationary linear process \(X\) with spectral density function \[ f(x)=| 1-e^{i(x-\omega_0)}|^{-d}| 1-e^{i(x+\omega_0)}|^{-d}f^{*}(x), \quad x\in[-\pi,\pi]\setminus \{\pm\omega_0\}, \] where \(0<d<1/2\) if \(\omega_0=0\) or \(\omega_0=\pi\), and \(0<d<1\) if \(\omega_0\in(0,\pi)\), and \(f^{*}(x)\) is a positive and continuous function on \([-\pi,\pi]\). The authors propose a consistent semi-parametric estimator of \(\omega_0\) with the best possible rate of convergence. They investigate consequences that the lack of knowledge of \(\omega_0\) might have on the estimation of \[ \alpha=\begin{cases} 2d & \text{if \(\omega_0\in\{0,\pi\}\)},\\ d& \text{if \(\omega_0\in(0,\pi)\)}.\end{cases} \] The authors show that a modified version of the \textit{J. Geweke} and \textit{S. Porter-Hudak} [ibid. 4, 221--238 (1983; Zbl 0534.62062)] estimator has the same rate-optimality properties when \(\omega_0\) is known, and it is asymptotically Gaussian with the same asymptotic variance. That is, statistical properties of this estimator of \(\alpha\) are not affected by the lack of knowledge of \(\omega_0\). A short Monte-Carlo experiment confirms these theoretical results.
      0 references

      Identifiers