On the estimation of coefficients of a simultaneous linear explosive model of higher orders with moving average errors generating a pair of time series (Q3979548)

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 20143
Language Label Description Also known as
default for all languages
No label defined
    English
    On the estimation of coefficients of a simultaneous linear explosive model of higher orders with moving average errors generating a pair of time series
    scientific article; zbMATH DE number 20143

      Statements

      On the estimation of coefficients of a simultaneous linear explosive model of higher orders with moving average errors generating a pair of time series (English)
      0 references
      26 June 1992
      0 references
      autoregressive model
      0 references
      Yule-Walker type estimators
      0 references
      CAN-estimators
      0 references
      partially explosive
      0 references
      roots
      0 references
      characteristic polynomial
      0 references
      unit circle
      0 references
      moving average residuals
      0 references
      moving average errors
      0 references
      two-stage estimation procedure
      0 references
      consistent and asymptotically normal estimators
      0 references
      simplified linear equations
      0 references
      Yule-Walker type equations
      0 references
      The model giving rise to the title of the paper is \[ X(t+2)-c_ 1 X(t+1)-c_ 2 X(t)-c_ 3 Y(t+1)-c_ 4 Y(t)-c_ 5=e_ 1(t+2), \] \[ Y(t+2)-d_ 1 Y(t+1)-d_ 2 Y(t)-d_ 3 X(t+1)-d_ 4 X(t)-d_ 5=e_ 2(t+2). \] The model is ``partially explosive'' in that some of the roots of the common characteristic polynomial lie inside and some outside the unit circle. The model with moving average residuals is considered. Least squares estimators are not consistent due to the presence of moving average errors; Yule-Walker estimators are not consistent due to the explosive nature of the model. In view of this, the authors adopt a two- stage estimation procedure that provides consistent and asymptotically normal estimators for the coefficients of the model. In this procedure, they provide two alternative versions of the first stage. Proofs and discussions are provided, but no numerical illustrations. A basic tool in the analysis is a re-presentation of the model as a set of simplified linear equations. The first stage of the estimation procedure consists of minimizing certain sums of squares; the second stage consists of posing a set of Yule-Walker type equations.
      0 references
      0 references
      0 references

      Identifiers