Estimating the Hurst parameter in fractional ARIMA(p,d,q) models via the quasi-likelihood method

This paper is concerned with R/S analysis given a fractional ARIMA(p,d,q) model with finite variance where the aim is to estimate the intensity of long-range dependence of the particular series. This is done through what is commonly referred to as the Hurst parameter (denoted by H). H is a measure of self-similarity of a given time series. The goal of this paper is to examine the effectiveness of applying the method of asymptotic quasi-likelihood to R/S analysis instead of the conventional method of least squares.