Hellinger transform of Gaussian autoregressive processes

DOI10.1016/0898-1221(94)90145-7MaRDI QIDQ1324401zbMATH OpenOpenAlexFDO

Authors István Fazekas

Publication date 17 November 1994

Published in Computers & Mathematics with Applications (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1016/0898-1221(94)90145-7

zbMATH Keywords

autoregressive processes Hellinger transform spectral densities

Mathematics Subject Classification ID

Stationary stochastic processes (60G10)

Review text

Let \(P^{(t)}\) denote the distribution of \(\{y_ 0,\ldots,y_ t\}\) where \(\{y_ t\}\) is an \(m\)-dimensional stationary Gaussian AR(1) process with zero mean, \(y_{t+1}=Ay_ t+\varepsilon_{t+1}\). Denote by \(P_ i^{(t)}\), \(1\leq i\leq N\), corresponding distributions if \(A=A_ i\). Let \[ H_ \alpha(P_ 1^{(t)},\ldots,P_ N^{(t)})=\int \bigl(dP_ 1^{(t)}\bigr)^{\alpha_ 1}\cdots \bigl( dP_ N^{(t)} \bigr)^{\alpha_ N},\qquad 0<\alpha_ i<1,\qquad \sum^ N_{i=1}\alpha_ i=1, \] denote the Hellinger transform. The Hellinger transform tends exponentially to zero as \(t\to\infty\). The main result of the paper is an explicit expression for this rate, i.e. \(\lim_{t\to\infty}(1/(t+1))\ln H_ \alpha(P_ 1^{(t)},\ldots,P_ N^{(t)})\) is explicitly evaluated in terms of the corresponding spectral densities. This result is sharpened for one-dimensional autoregressive processes.

Recommendations

Cites work

Cited in

(4)

This page was built for publication: Hellinger transform of Gaussian autoregressive processes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1324401)