Consistent Estimates of Autoregressive Parameters and Extended Sample Autocorrelation Function for Stationary and Nonstationary ARMA Models (Q3323072)

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 3854245
Language Label Description Also known as
default for all languages
No label defined
    English
    Consistent Estimates of Autoregressive Parameters and Extended Sample Autocorrelation Function for Stationary and Nonstationary ARMA Models
    scientific article; zbMATH DE number 3854245

      Statements

      0 references
      0 references
      1984
      0 references
      unified approach
      0 references
      time series
      0 references
      AIC
      0 references
      ARMA models
      0 references
      extended sample autocorrelations
      0 references
      S-array
      0 references
      Yule-Walker equations
      0 references
      backshift operator
      0 references
      Gaussian white noise process
      0 references
      iterative regression procedure
      0 references
      consistent least squares estimates
      0 references
      Consistent Estimates of Autoregressive Parameters and Extended Sample Autocorrelation Function for Stationary and Nonstationary ARMA Models (English)
      0 references
      In this paper the authors consider the ARMA(p,q) model for a time series \(\{Z_ t\}: \Phi(B)Z_ t=C+\theta(B)a_ t\), where \(\Phi(B)=U(B)\phi(B)=1-\Phi_ 1B-\cdot \cdot \cdot -\Phi_ pB^ p\), \(U(B)=1-U_ 1B-\cdot \cdot \cdot -U_ dB^ d\), \(\phi(B)=1-\phi_ 1B- \cdot \cdot \cdot -\phi_{p-d}B^{p-d}\), \(\theta(B)=1-\theta_ 1B- \cdot \cdot \cdot -\theta_ qB^ q\) are polynomials in B, B is backshift operator such that \(BZ_ t=Z_{t-1}\), C is a constant; \(\{a_ t\}\) is a Gaussian white noise process, \(Ea_ t=0\), \(Ea^ 2_ t=\sigma^ 2_ a.\) The authors propose an iterative regression procedure that yields consistent least squares estimates for the autoregressive parameters. It is shown that the least squares estimates \({\hat \Phi}^{(j)}_{l(d)}\) for any j-th iterated AR(d) regression are consistent for the nonstationary autoregressive coefficients \(U_ l\). Properties of the iterated autoregressive estimates are studied. The problem of tentative model identification is considered also.
      0 references

      Identifiers