scientific article; zbMATH DE number 3982353

From MaRDI portal
Publication:3746737





The statistical analysis of many applied problems like an investigation of hydrological time series and others has led to the conclusion that the peak of the periodogram near the origin should be explained by a model with spectral density which is not bounded in the neighbourhood of zero frequency. For this reason the paper investigates time series models with a long memory, namely, a stationary process X(t) with covariance function \(R_ k,\sum | R_ k| =\infty\) is considered. As an example of such a process the paper investigates the fractionally differenced white noise \(X_ t\) satisfying \[ (1-B)^{\delta}X_ t=\epsilon_ t,\quad \delta \in (-1/2,1/2) \] where \(\epsilon_ t\) is a white noise, B is a back-shift operator \(BX_ t=X_{t-1}\); as well as the seasonal fractionally differenced white noise defined by \[ (1-B^ s)^{\delta}X_ t=\epsilon_ t \] or the seasonal persistent process \[ [(1-e^{i\omega}B)(1-e^{-i\omega}B)]^{\delta}X_ t=\epsilon_ t. \] Spectral densities, correlation functions, estimations of parameters are investigated.




Cited in
(56)








This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3746737)