Methods of Signal Processing

Nonparametric Signal Estimation when There Is Incomplete Information on the Noise Distribution

A. B. Tsybakov


Abstract: The article proposes a class of nonparametric signal estimates on the basis of observations in noise with probability characteristics that are not precisely known. The consistency conditions for these estimates are given, and their rate of convergence is determined. The problem of choosing estimates that are optimal in the sense of rate of convergence is solved.

UDC: 621.391.1:519.28

Received: 15.06.1981

 English version: Problems of Information Transmission, 1982, 18:2, 116–130

Bibliographic databases:

• 
•