Fractional Time Series Modelling

Aspects of model building using fractionally differenced autoregressive-moving average processes are discussed. An algorithm for approximate maximum likelihood estimation is outlined and the large-sample distribution of the maximum likelihood estimates is derived. The large-sample distribution of th...

Description complète

Détails bibliographiques
Publié dans:Biometrika. - Oxford University Press. - 73(1986), 1, Seite 217-221
Auteur principal: Li, W. K. (Auteur)
Autres auteurs: McLeod, A. I.
Format: Article en ligne
Langue:English
Publié: 1986
Accès à la collection:Biometrika
Sujets:Fractional differencing Long-memory time series Maximum likelihood Parsimony Portmanteau test Residual autocorrelation
Description
Résumé:Aspects of model building using fractionally differenced autoregressive-moving average processes are discussed. An algorithm for approximate maximum likelihood estimation is outlined and the large-sample distribution of the maximum likelihood estimates is derived. The large-sample distribution of the residual autocorrelations is also derived and a modified portmanteau test statistic is obtained for checking model adequacy.
ISSN:14643510
DOI:10.2307/2336289