Abstract:
In this article tire seasonal autoregressive moving-average (ARMA) model is compared with the periodic ARMA (PARMA) model via their autocorrelation functions (ACF). Some low order models of either type are considered and their theoretical ACF are provided. General formulas are observed for the computation of the theoretical ACF of various PARMA models. Some specific examples are given for both types of models and their ACF`s are sketched. While any seasonal ARMA model has a single ACF, it is observed that, in general, any PARMA model does not have a unique ACF, it may rather has as many ACF`s as the period length. Thus, the ordinary ACF, can not be a good tool for the identification of PARMA models.
Page(s):
199-212
DOI:
DOI not available
Published:
Journal: Pakistan Journal of Statistics, Volume: 20, Issue: 2, Year: 2004