A test for randomness against ARMA alternatives
In a recent paper, Mokkadem (1997. Stoch. Proc. Appl. 72, 145-149) derived a simple test for randomness against ARMA alternatives. In this note we consider a transformation of the corresponding statistic and present an alternative proof of this result. Through this approach it is demonstrated that the asymptotic distribution of the corresponding test statistic under the alternative does not depend on the fourth-order moments of the innovations. A simulation study indicates that a transformation of the test of Mokkadem (1997) (which is asymptotically equivalent to the classical portmanteau test) may perform better for finite sample sizes.
Year of publication: |
2000
|
---|---|
Authors: | Dette, Holger ; Spreckelsen, Ingrid |
Published in: |
Stochastic Processes and their Applications. - Elsevier, ISSN 0304-4149. - Vol. 89.2000, 1, p. 131-139
|
Publisher: |
Elsevier |
Keywords: | Time series Test for randomness ARMA process |
Saved in:
Saved in favorites
Similar items by person
-
Some comments on specification tests in nonparametric absolutely regular processes
Dette, Holger, (2001)
-
A note on a specification test for time series models based on spectral density estimation
Dette, Holger, (2001)
-
A note on a specification test for time series models based on spectral density estimation
Dette, Holger, (2001)
- More ...