A possible definition of a stationary tangent

U. Keich

Stochastic Processes and their Applications, 2000, vol. 88, issue 1, 1-36

Abstract: This paper offers a way to construct a locally optimal stationary approximation for a non-stationary Gaussian process. In cases where this construction leads to a unique stationary approximation we call it a stationary tangent. This is the case with Gaussian processes governed by smooth n-dimensional correlations. We associate these correlations with equivalence classes of curves in . These are described in terms of "curvatures" (closely related to the classical curvature functions); they are constant if and only if the correlation is stationary. Thus, the stationary tangent, at t=t0, to a smooth correlation, curve or process, is the one with the same curvatures at t0 (but constant). We show that the curvatures measure the quality of a local stationary approximation and that the tangent is optimal in this regard. These results extend to the smooth infinite-dimensional case although, since the equivalence between correlations and curvatures breaks down in the infinite-dimensional setting, we cannot, in general, single out a unique tangent. The question of existence and uniqueness of a stationary process with given curvatures is intimately related with the classical moment problem and is studied here by using tools from operator theory. In particular, we find that there always exists an optimal Gaussian approximation (defined via the curvatures). Finally, by way of discretizing we introduce the notion of [delta]-curvatures designed to address non-smooth correlations.

Keywords: Stationary; processes; Curvature; functions; Stationary; tangent; Moment; problem; Tri-diagonal; matrices (search for similar items in EconPapers)
Date: 2000
References: View complete reference list from CitEc
Citations: View citations in EconPapers (1)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0304-4149(99)00116-7
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:spapps:v:88:y:2000:i:1:p:1-36

Ordering information: This journal article can be ordered from
http://http://www.elsevier.com/wps/find/supportfaq.cws_home/regional
https://shop.elsevie ... _01_ooc_1&version=01

Access Statistics for this article

Stochastic Processes and their Applications is currently edited by T. Mikosch

More articles in Stochastic Processes and their Applications from Elsevier
Bibliographic data for series maintained by Catherine Liu ().