Fotis Papailias

fotis-papailias-cv.pdf

Essays on Long Memory Time Series

My research focuses on topics in long-range dependent time series. During the last decades, fractional processes have become very popular due to their wide applicability in physics, geology, macroeconomics, finance etc. The newly introduced type of models that combine long memory with nonlinear characteristics (FI(d)-NLAR(p,k)-ESTAR) is analyzed. The general ARFIMA model is used, when its zero integrated part follows an exponential STAR model. I am specifically interested in the investigation of the parameter estimates, the autocorrelations and the impulse responses of such series via Monte Carlo Study. Furthermore, the use of block bootstraps in such series has been of interest to me. A number of simulation experiments in different models shows that this method can provide better results than its competitor in such models.