An empirical central limit theorem in L\(^1\) for stationary sequences (Q734650): Difference between revisions
From MaRDI portal
Set OpenAlex properties. |
Changed an Item |
||
Property / arXiv ID | |||
Property / arXiv ID: 0812.2839 / rank | |||
Normal rank |
Revision as of 17:35, 18 April 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | An empirical central limit theorem in L\(^1\) for stationary sequences |
scientific article |
Statements
An empirical central limit theorem in L\(^1\) for stationary sequences (English)
0 references
13 October 2009
0 references
Let \(\mu\) be a \(\sigma\)-finite measure, \(\mathbb{T}\) be a real interval and \(\mathbb{L}^2(\mu)= \mathbb{L}^1(\mathbb{T},\mu)\). Let \((X_i)_{i\in\mathbb{Z}}\) be an ergodic stationary sequence of centered random variables in \(\mathbb{L}^1(\mu)\) such that \(X_i= X_0 T^i\), where \(T: \Omega\to\Omega\) is a bijective bi-measurable transformation preserving the probability \(\mathbb{P}\) on \((\Omega,{\mathcal A})\). For a \(\sigma\)-algebra \({\mathcal F}_0\) of \({\mathcal A}\), satisfying \({\mathcal F}_0\subset T^{-1}({\mathcal F}_0)\), let \({\mathcal F}_i= T^{-i}({\mathcal F}_0)\), \({\mathcal F}_\infty= \bigcap_{n\geq 0}{\mathcal F}_{-n}\) and \({\mathcal F}_\infty= \bigvee_{k\in\mathbb{Z}}{\mathcal F}_k\). Let \(P_0(X(t))= E(X_0(t)|{\mathcal F}_0)- E(X_0(t)|{\mathcal F}_{-1})\). In this paper, the author proves the following central limit theorem: Assume that, for any real \(t\), \(E(X_0(t)|{\mathcal F}_{-\infty})= 0\), \(E(X_0(t)|{\mathcal F}_\infty)= X_0(t)\) and \[ \sum_{k\in\mathbb{Z}}\,\int_{\mathbb{T}}\| P_0(X_k(t))\|_2\mu(dt)< \infty. \] Then \[ n^{-1/2} \sum^n_{i=1} X_0\circ T^i\to G\text{ in law in }\mathbb{L}^1(\mu), \] where \(G\) is an \(\mathbb{L}^1(\mu)\)-valued centered Gaussian random variables with covariance operator such that for any \((f,g)\) in \(\mathbb{L}^\infty(\mu)\times\mathbb{L}^\infty(\mu)\), \[ E\Biggl(f\Biggl(\sum_{k\in\mathbb{Z}} P_0(X_k)\Biggr) g\Biggl(\sum_{k\in\mathbb{Z}} P_0(X_k)\Biggr)\Biggr)= \sum_{k\in\mathbb{Z}} \text{Cov}(f(X_0), g(X_k)). \] Via Corollary to this theorem the author also derives sufficient conditions for the convergence in \(\mathbb{L}^1\) of the empirical process defined by a stationary sequence \((Y_i)\) of real-valued random variables as soon as \((Y_i)\) satisfies some weak dependence conditions.
0 references
empirical distribution function
0 references
central limit theorem
0 references
ergodic stationary sequence
0 references
Wasserstein distance
0 references