A central limit theorem for isotropic random walks on \(n\)-spheres for \(n\to \infty\) (Q1842049)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | A central limit theorem for isotropic random walks on \(n\)-spheres for \(n\to \infty\) |
scientific article |
Statements
A central limit theorem for isotropic random walks on \(n\)-spheres for \(n\to \infty\) (English)
0 references
5 October 1995
0 references
Let \((Y_ l)_{l \geq 0}\) be an isotropic, time-homogeneous Markov chain on \(S^ n\) starting at a fixed point \(x \in S^ n\), satisfying \(P(\cos \angle (Y_ l, x) = r \mid Y_{l-1} = x) =1\) for all \(l\) and \(x\). The following central limit theorem for \(X^ n_ l = \cos \angle (Y_ l, x_ 0)\) is established. Let \(l, n \to \infty\) and \(r \to 1\) in such a manner: \(n^ \rho (1-r) \to 0\), \(\rho > 0\); \(l(1-t) - {1\over 2} \ln n \to b\in (-\infty,\infty]\). Then \(\sqrt {n} X^ n_ l\) converges in distribution to \(N(e^{-b},1)\). The proof is given for a more general setting of Markov chains on \([-1,1]\) with convolution structure by using of limiting properties of ultraspherical polynomials relating to Hermite polynomials and the moment convergence theorem. Further applications of this method of proof are discussed.
0 references
isotropic random walk
0 references
central limit theorem
0 references
Hermite polynomials
0 references
moment convergence theorem
0 references