scientific article; zbMATH DE number 3911540

From MaRDI portal
Publication:3687560





[For the entire collection see Zbl 0553.00008.] A real valued stationary time series \(\{\) X(t)\(\}\), \(t\in Z\), with mean 0 and spectral density \(f_{\theta}(\lambda)\), depending on the unknown vector parameter \(\theta =(\theta_ 1,\theta_ 2,...,\theta_ p)\) is considered and the problem of statistical estimation of the unknown parameter \(\theta\) is studied. In the case of stationary Gaussian processes the likelihood function can be presented explicitly, however the realization of this method of unknown parameters estimation is practically difficult. The author considers the Whittle modification of the likelihood function in which instead of the ordinary periodogram the so called tapered periodogram is taken. Consistency and asymptotic normality of the parameter estimator are proved and the problem is investigated what kind of tapers lead to asymptotically efficient estimates.











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 Q3687560)