Abstract:
We consider the problem of recovery (minimax estimation) of a linear functional under random errors in observations. An estimate of the loss of quality, due to employing linear methods of recovery, is obtained. A new formalization is also proposed for an informative problem of recovery under random errors in information which permits, in various cases, an effective determination of optimal solution without passing to asymptotics in noise.
Keywords:minimax estimation, recovery of a linear functional.