CMStatistics 2018: Start Registration
View Submission - CFE
A0234
Title: Autoregressive tempered fractionally integrated moving average time series: Theory and applications Authors:  Farzad Sabzikar - Iowa State University (United States) [presenting]
Abstract: The autoregressive tempered fractionally integrated moving average (ARTFIMA) time series model applies a tempered fractional difference to the standard ARMA time series. The ARTFIMA model can also be interpreted as an extension of the ARFIMA model. ARTFIMA time series exhibit semi-long range dependence: Their covariance function resembles long range dependence for a number of lags, depending on the tempering parameter, but eventually decays exponentially fast. The mathematical foundation for ARTFIMA parameter estimation will be discussed. A new R package artfima to fit data will be presented. Several examples from finance, geophysics, turbulence, and climate illustrate the fitting procedure, and the utility of the ARTFIMA model. Finally, an invariance principles for ARTFIMA times series will be given when the tempering parameter depends on the sample size.