Approximating martingales and the central limit theorem for strictly stationary processes (Q1208930): Difference between revisions

Consider a space \((\Omega,{\mathcal A},\mu)\) and a measure preserving transformation \(T:\Omega\to\Omega\). Let \(f\in L_ 2(\mu)\) and define \(s_ n(f)={1\over\sqrt n}\sum^ n_{j=1}f\circ T^ j\). The author studies assumptions of various central limit theorems (like those of Gordin and Dürr-Goldstein) which ensure that \(s_ n(f)\) weakly converges to a normal law. The theorems can be characterized by the fact that they guarantee \(\lim_{n\to\infty}\| s_ n(f-m)\|_ 2=0\) for any martingale difference \(m\circ T^ i\). Then the author discusses conditions under which the results hold in almost all ergodic components simultaneously. An application to the asymptotic theory of stationary linear processes with random coefficient is given.