Asymptotic expansions for functions of the increments of certain Gaussian processes (Q2267546): Difference between revisions

Let \(\{G(x), x\geq 0\}\) be a Gaussian process with zero mean and stationary increments and denote by \(\sigma^2(h)=\text{Var}\,G(h)\) the incremental variance of \(G\). Let \(f:\mathbb R\to\mathbb R\) be a function. Under appropriate conditions on \(\sigma^2\) and \(f\), the authors provide an \(L^2\)-asymptotic expansion for the integral \[ I_{[a,b]}(h):=\int_{a}^b f\left(\frac{G(x+h)-G(x)}{\sigma(h)}\right)dx \] as \(h\downarrow 0\) in terms of the Wick powers of the generalized derivative of \(G\). As a consequence of this result, central limit theorems with normal or non-normal limit distribution are obtained for the integral \(I_{[a,b]}(h)\) as \(h\downarrow 0\). Central limit theorems for non-linear functionals of stationary long-range dependent Gaussian processes have been a topic of active research, see for example [\textit{M. S. Taqqu}, Z.\ Wahrscheinlichkeitstheor.\ Verw.\ Geb.\ 31, 287--302 (1975; Zbl 0303.60033), Z. Wahrscheinlichkeitstheor.\ Verw.\ Geb.\ 50, 53--83 (1979, Zbl 0397.60028)]. As opposed to these papers which consider \textit{global} non-linear functionals, the paper under review is concerned with the \textit{local} properties of Gaussian processes.