A Gaussian Radon transform for Banach spaces (Q694771): Difference between revisions

Let \(B\) be a real separable Banach space. The authors construct a Gaussian measure \(\mu_P\) for hyperplanes \(P\) in \(B\) and define a Gaussian Radon transform \(Gf(P)= \int f\,d\mu_P\) on functions \(f\) on \(B\). As they write, this construction is ``motivated by the task of recovering information about a random variable \(f\), such as a Brownian functional, from certain conditional expectations of \(f\)''. The main result of the paper is a uniqueness type theorem: If \(f\) is a bounded continuous function on the real separable Banach space \(B\) which is the completion of a real separable Hilbert space \(H\) with respect to a measurable norm, and if \(Gf= 0\) on all hyperplanes of \(H\) which do not intersect a given closed, bounded, convex subset \(K\subset H\), then \(f=0\) on the complement of \(K\) in \(B\).