On the asymptotic distributions of partial sums of functionals of infinite-variance moving averages (Q1568300): Difference between revisions

Assume \(\{\varepsilon_n\}\) is an i.i.d. sequence of random variables such that \(E\exp(it\varepsilon_n)= \exp(-|t|^\alpha)\), where \(\alpha\in (0,2)\). Let \(\{a_n\}\) be a sequence of constants for which the linear process \(X_n= \sum^\infty_{j=1} a_j\varepsilon_{n- j}\), \(n\geq 1\), is well defined. Let \(K\) be a measurable and bounded function. The author investigates the asymptotic distribution of the partial sums \(S_N= \sum^N_{n=1} [K(X_n)- EK(X_n)]\) as \(N\to\infty\). The presented results show that \(S_N\) follows a central or noncentral limit theorem depending on the rate at which the moving average coefficients tend to \(0\).