Estimation of a linear functional from indirect observations
Let X be a Wiener process with intensity \(\epsilon\) and drift S which is a square-integrable function belonging to a certain admissible set. The observation process Y is assumed to satisfy an Itô equation \(dY=X dt+\delta dW\), where W is a Wiener process independent of X. The drift S is unknown and the problem is to estimate a certain linear functional of S given the observation Y. An estimator is said to be linear if it is of the form \(\int^{t}_{0}m dY\), where m is a square-integrable function. The problem considered is to construct an estimator which is minimax in the class of linear estimators with respect to a quadratic loss function. Under some assumptions on the admissible set the general formulas for the minimax linear estimator and for the value of the minimax risk are derived. A special case, when the admissible set is a, possibly degenerated, ellipsoid, is discussed in detail. Explicit formulas for the estimator and minimax risk are given. It is shown that in some cases, for small intensities \(\epsilon\) and \(\delta\), the minimax risk is approximately the same as in the case of the process X directly observed.
This page was built for publication: Estimation of a linear functional from indirect observations
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1099562)
