Rescaled range analysis in the presence of stochastic trend
We study the limiting behavior of the prominent R/S test statistic, aimed at detecting long-range dependence, if instead of long memory a stochastic trend given by cumulative random shocks is present. As the main result we derive the convergence rate of the R/S statistic to its limit.
Year of publication: |
2007
|
---|---|
Authors: | Aue, Alexander ; Horváth, Lajos ; Steinebach, Josef |
Published in: |
Statistics & Probability Letters. - Elsevier, ISSN 0167-7152. - Vol. 77.2007, 12, p. 1165-1175
|
Publisher: |
Elsevier |
Keywords: | Compound Poisson process Long memory Long-range dependence Rescaled range analysis Rescaled variance analysis R/S statistic Stochastic trend V/S statistic |
Saved in:
Saved in favorites
Similar items by person
-
Monitoring shifts in mean: Asymptotic normality of stopping times
Aue, Alexander, (2008)
-
Estimation in Random Coefficient Autoregressive Models
Aue, Alexander, (2006)
-
Sequential testing for the stability of high-frequency portfolio betas
Aue, Alexander, (2012)
- More ...