We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math.ST

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Statistics Theory

Title: On minimax nonparametric estimation of signal in Gaussian noise

Abstract: For the problem of nonparametric estimation of signal in Gaussian noise we point out the strong asymptotically minimax estimators on maxisets for linear estimators (see \cite{ker93,rio}). It turns out that the order of rates of convergence of Pinsker estimator on this maxisets is worse than the order of rates of convergence for the class of linear estimators considered on this maxisets. We show that balls in Sobolev spaces are maxisets for Pinsker estimators.
Subjects: Statistics Theory (math.ST)
MSC classes: 65M30, 65R30, 62G08, 62J07
Cite as: arXiv:1710.00576 [math.ST]
  (or arXiv:1710.00576v2 [math.ST] for this version)

Submission history

From: Mikhail Ermakov s [view email]
[v1] Mon, 2 Oct 2017 10:56:35 GMT (8kb)
[v2] Sun, 5 Nov 2017 11:38:56 GMT (9kb)

Link back to: arXiv, form interface, contact.