Divergence of an integral of a process with small ball estimate

The problem of convergence/divergence of integral functionals \(\int_{0}^{\infty}g(X(t))dt\) for several calasses stochastic processes and several classes of functions \(g\) usually appears quite often, particularly in some statistical applications. Let \(X=\{X(t),t\geq 0\}\) be a one-dimensional stochastic process with continuous trajectories, and let \(g:\mathbb{R}\rightarrow \mathbb{R}\) be a continuous function. Then for any \(T>0\) the integral functional \(\int_{0}^{T} g(X(t)) dt\) is defined. However, its properties and its asymptotic behaviour as \(T\rightarrow \infty\) depend on the properties of the process \(X\) and the function \(g\). The asymptotic behaviour of the integral functional \(\int_{0}^{T} g(X(t)) dt\) is very different even for one-dimensional Markov processes and depends on their transient or recurrent properties. With these considerations the authors provide sufficient conditions on the function \(f\) and the stochastic process \(X\) that supply the divergence of the integral functional \(\int_{0}^{T} f(X(t))^2 dt\) at the rate \(T^{1-\varepsilon}\) as \(T\rightarrow \infty\) for \(\varepsilon >0\). The given conditions include the so called small ball estimates which are given and discussed in detail. Some statistical applications are also given.

Recommendations

Cites work

Cited in

(1)

Asymptotic covariances for functionals of weakly stationary random fields

This page was built for publication: Divergence of an integral of a process with small ball estimate

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2132525)