Asymptotics for fractional processes

From MaRDI portal





``Asymptotics for fractional processes, by James Davidson, investigates the large sample behaviour of fractional partial-sum processes involving long memory increments. Long memory or long range dependence in a time series is defined to mean that the sum of the auto covariance of the time series is divergent. The processes discussed here have a linear moving average representation with a parameter denoted by \(d\). If \(d\) is positive, then the time series is said to have long memory and the number \(d\) measures the degree of long-run persistence and if \(d\) is negative, then the time series is said to have a short memory known as antipersistence. Chapters 1, 2 and 3 deal with the weak convergence of certain normalized partial sums of the time series to the fractional Brownian motion which is an almost surely continuous Gaussian process with correlated increments. Chapters 4, 5 and 6 investigate the limiting distribution of stochastic integrals where both the integrand and integrator processes exhibit either long memory or antipersistence as described above. Chapter 7 reviews applications of the theory to regression with fractional processes. Chapter 8 deals with extensions of the results to driving forces or shocks which are dependent. Some related results on frequency domain analysis and autoregressive roots near unity are presented in Chapters 9 and 10. The results are well presented in a theorem-proof style and I recommend the book to any reader interested in the study of asymptotics for fractional processes.












This page was built for publication: Asymptotics for fractional processes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7025103)