The functional central limit theorem and weak convergence to stochastic integrals. II: Fractionally integrated processes (Q2716482)

The authors continue with functional central limit results obtained in part I (reviewed above) for a broad class of serially dependent and heterogeneously distributed vector processes, which are weakly dependent, otherwise characterized as ``short memory processes''. The present paper extends previous results by allowing the processes to exhibit long memory. The authors consider fractionally integrated processes or, in other words, \(I(d)\) processes, with \(|d|<1\), and the limit processes for their sums are fractional Brownian motions. The prelimit variables may exhibit quite general weak dependence by being near-epoch-dependent functions of mixing processes. Multivariate extensions of FCLT are presented. Some weak convergence results for stochastic integrals having fractional integrands and weakly dependent integrators are also obtained. These theorems need different approaches for negative and positive values of \(d\).