Fractionally Integrated Time Series Analysis

This section describes subroutines related to fractionally integrated time series analysis. The phenomenon of long memory can be observed in hydrology, finance, economics, and so on. Unlike with a stationary process, the correlations between observations of a long memory series are slowly decaying to zero.

The following subroutines are supported:

FARMACOV

computes the autocovariance function for a fractionally integrated ARMA model.

FARMAFIT

estimates the parameters for a fractionally integrated ARMA model.

FARMALIK

computes the log-likelihood function for a fractionally integrated ARMA model.

FARMASIM

generates a fractionally integrated ARMA process.

FDIF

computes a fractionally differenced process.