Abstract.
The paper deals with the decomposition of a time series process admitting an ARIMA representation into permanent and transitory components, with the intent of investigating whether the introduction of correlated disturbances provides meaningful extensions of the admissible parameter range. The main points are illustrated with reference to ARIMA(2,1,0) and IMA(2,2) models. It is argued that there is very little reason for such extensions, and that the restrictions implied by the assumption of uncorrelated components are sound.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Author information
Authors and Affiliations
Corresponding author
Additional information
This research was supported by the MURST Cofin2000. The paper was presented at the XL Scientific Meeting of the Italian Statistical Society (Florence 2000), ISF 2000 (Lisbon), and at the Europaeisches Heimbildungswerk workshop, Helenau-Bernau (Berlin). I thank participants for their comments and in particular Jörg Breitung for very stimulating discussion. I also wish to thank the associate editor and the referee for their comments.
Rights and permissions
About this article
Cite this article
Proietti, T. Unobserved components models with correlated disturbances. Statistical Methods & Applications 12, 277–292 (2004). https://doi.org/10.1007/s10260-003-0074-y
Issue Date:
DOI: https://doi.org/10.1007/s10260-003-0074-y