Two remarks on normality preserving Borel automorphisms of R^n

From MaRDI portal
Publication:1948992



Abstract: Let T be a bijective map on mathbbRn such that both T and T−1 are Borel measurable. For any and any real nimesn positive definite matrix Sigma, let denote the n-variate normal (gaussian) probability measure on mathbbRn with mean vector and covariance matrix Sigma. Here we prove the following two results: (1) Suppose is gaussian for 0leqjleqn where I is the identity matrix and is a basis for mathbbRn. Then T is an affine linear transformation; (2) Let Sigmaj=I+epsilonjmathbfujmathbfujprime, 1leqjleqn where epsilonj>−1 for every j and mathbfuj,1leqjleqn is a basis of unit vectors in mathbbRn with mathbfujprime denoting the transpose of the column vector mathbfuj. Suppose N(mathbf0,I)T−1 and N(mathbf0,Sigmaj)T−1, 1leqjleqn are gaussian. Then T(mathbfx)=sumlimitsmathbfs1EmathbfsVmathbfsUmathbfx a.e. mathbfx where mathbfs runs over the set of 2n diagonal matrices of order n with diagonal entries pm1, U,,V are nimesn orthogonal matrices and Emathbfs is a collection of 2n Borel subsets of mathbbRn such that Emathbfs and VmathbfsU(Emathbfs) are partitions of mathbbRn modulo Lebesgue-null sets and for every j, VmathbfsUSigmaj(VmathbfsU)−1 is independent of all mathbfs for which the Lebesgue measure of Emathbfs is positive. The converse of this result also holds. vskip0.1in Our results constitute a sharpening of the results of S. Nabeya and T. Kariya












This page was built for publication: Two remarks on normality preserving Borel automorphisms of \(\mathbb{R}^{n}\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1948992)