A characterization of Gaussian measures

From MaRDI portal





Let S denote the class of all symmetric probability measures \(\mu\) on the real line. Then the author calls \(\mu\in S\) weakly stable if for independent random variables \(X_ 1,X_ 2\sim \mu\) and positive numbers \(a_ 1,a_ 2\) there exists a non-negative random variable Y such that \(a_ 1X_ 1+a_ 2X_ 2\sim YX\), where \(X\sim \mu\) and X is independent of Y. The following theorems are proved: Theorem 1: A measure \(\mu\sim S\) is weakly stable, infinitely divisible and satisfies \(\int | x|^ p\mu(dx)<\infty\) for all p, \(0<p<2\), iff it is Gaussian. Theorem 2: A measure \(\mu\in S\) is weakly stable and satisfies \(\int | x|^ p\mu(dx)<\infty\) for all p, \(0<p<2\), and \(\int \exp \{| x|^{2+\epsilon}\}\mu(dx)=\infty\) for all \(\epsilon>0\) iff it is non-degenerate Gaussian. Theorem 3: A measure \(\mu\in S\) is weakly stable with at least one atom iff it is a symmetric two-point distribution.











This page was built for publication: A characterization of Gaussian measures

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3319479)