English

Martingale approximation and optimality of some conditions for the central limit theorem

Probability 2009-12-16 v1

Abstract

Let (Xi)(X_i) be a stationary and ergodic Markov chain with kernel QQ, ff an L2L^2 function on its state space. If QQ is a normal operator and f=(IQ)1/2gf = (I-Q)^{1/2}g (which is equivalent to the convergence of n=1k=0n1Qkfn3/2\sum_{n=1}^\infty \frac{\sum_{k=0}^{n-1}Q^kf}{n^{3/2}} in L2L^2), we have the central limit theorem (cf\. \cite{D-L 1}, \cite{G-L 2}). Without assuming normality of QQ, the CLT is implied by the convergence of n=1k=0n1Qkf2n3/2\sum_{n=1}^\infty \frac{\|\sum_{k=0}^{n-1}Q^kf\|_2}{n^{3/2}}, in particular by k=0n1Qkf2=o(n/logqn)\|\sum_{k=0}^{n-1}Q^kf\|_2 = o(\sqrt n/\log^q n), q>1q>1 by \cite{M-Wu} and \cite{Wu-Wo} respectively. We shall show that if QQ is not normal and f(IQ)1/2L2f\in (I-Q)^{1/2} L^2, or if the conditions of Maxwell and Woodroofe or of Wu and Woodroofe are weakened to n=1cnk=0n1Qkf2n3/2<\sum_{n=1}^\infty c_n\frac{\|\sum_{k=0}^{n-1}Q^kf\|_2}{n^{3/2}}<\infty for some sequence cn0c_n\searrow 0, or by k=0n1Qkf2=O(n/logn)\|\sum_{k=0}^{n-1}Q^kf\|_2 = O(\sqrt n/\log n), the CLT need not hold.

Keywords

Cite

@article{arxiv.0912.2864,
  title  = {Martingale approximation and optimality of some conditions for the central limit theorem},
  author = {Dalibor Volný},
  journal= {arXiv preprint arXiv:0912.2864},
  year   = {2009}
}

Comments

to appear in Journal of Theoretical Probability