An asymptotic property of nonlinear estimators arising as solutions to a certain class of convex programming problems (Q1083817)

The main question under consideration is how close are the conditional expectations E(Y\(| X(t):t\in [0,T])\) and \(E(Y| X(t_ i):1\leq i\leq n\), \(t_ i\in [0,T])\) to each other, where X(t) is a stochastic process and \(n\to \infty\). To answer this question the authors use some abstract Banach space principles and formulate theorems about convergence of the conditional expectations in terms of the Luxemburg norm in the Orlicz space.