English

An optimal linear filter for estimation of random functions in Hilbert space

Statistics Theory 2020-08-31 v1 Functional Analysis Statistics Theory

Abstract

Let {\mbox{\mbox{\boldmath ff}}} be a square-integrable, zero-mean, random vector with observable realizations in a Hilbert space HH, and let {\mbox{\mbox{\boldmath gg}}} be an associated square-integrable, zero-mean, random vector with realizations, which are not observable, in a Hilbert space KK. We seek an optimal filter in the form of a closed linear operator XX acting on the observable realizations of a proximate vector {\mbox{\mbox{\boldmath ff}}}_{\epsilon} \approx {\mbox{\mbox{\boldmath ff}}} that provides the best estimate \widehat{{\mbox{\mbox{\boldmath gg}}}}_{\epsilon} = X {\mbox{\mbox{\boldmath ff}}}_{\epsilon} of the vector {\mbox{\mbox{\boldmath ff}}}. We assume the required covariance operators are known. The results are illustrated with a typical example.

Keywords

Cite

@article{arxiv.2008.12485,
  title  = {An optimal linear filter for estimation of random functions in Hilbert space},
  author = {Phil Howlett and Anatoli Totokhti},
  journal= {arXiv preprint arXiv:2008.12485},
  year   = {2020}
}
R2 v1 2026-06-23T18:09:29.894Z