Description
Fitting and Forecasting Gegenbauer ARMA Time Series Models.
Description
Methods for estimating univariate long memory-seasonal/cyclical Gegenbauer time series processes. See for example (2022) <doi:10.1007/s00362-022-01290-3>. Refer to the vignette for details of fitting these processes.
README.md
garma - R package for estimation of Gegenbauer Seasonal/Cyclical long memory processes.
Overview & Introduction
This package fits a GARMA model (refer documentation) to a univariate time series.
GARMA models are extensions of ARIMA models which allow for both fractional differencing (like "fracdiff") but also allow that to happen at a non-zero frequency in the spectrum.
This package will estimate that frequency (which is known for technical reasons as the "Gegenbauer" frequency).
At time of writing several estimation methods are supports as well as a number of (non-linear) optimisation routines.
Installation.
The package can be installed from CRAN in the usual manner:
> install.packages('garma')
Documentation
An Introduction to the "garma" packages is available here, and the reference documentation is available here.