On the small balls problem for equivalent Gaussian measures

From MaRDI portal





The paper belongs to infinite-dimensional analysis. The following difficult problem is investigated. Let \(X\) be a locally convex space, \(\mu\) be a centred Gaussian Radon measure, \(\nu\) be another Gaussian Radon measure on \(X\) equivalent to \(\mu\) and \(q\) be an arbitrary \(\mu\)-measurable norm. Consider the fraction \[ J_q(\varepsilon)=\frac{\nu(x:q(x)\leq\varepsilon)}{\mu(x:q(x)\leq\varepsilon)} \] defined for \(\varepsilon>0\). From author's text: ``In this paper we study the problem of the existence of the limit \[ \lim_{\varepsilon \to 0}J_q(\varepsilon).\tag{0.1} \] The Radon-Nikodým density of \(\nu\) with respect to \(\mu\) has the form \(\exp Q\), where \(Q\) is a \(\mu\)-measurable second-order polynomial. The main result of this work is as follows: A positive limit in (0.1) exists for each \(\mu\)-measurable norm \(q\) on \(X\) precisely when the second derivative of \(Q\) along the Cameron-Martin space \(H=H(\mu)\dots\) is a nuclear operator on \(H\). The cases when the limit in (0.1) is zero, is infinite or does not exist at all are also completely described. The references to earlier related works are given (the bibliography contains 23 items). Surely, the author made a significant progress in the study of the considered problem.




Cited in
(only showing first 100 items - show all)








This page was built for publication: On the small balls problem for equivalent Gaussian measures

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4220890)