Estimation of Fractional ARIMA Process with Stable Innovations: A Monte Carlo StudyReport as inadecuate




Estimation of Fractional ARIMA Process with Stable Innovations: A Monte Carlo Study - Download this document for free, or read online. Document in PDF available to download.

1 Lerstad LERSTAD - laboratoire d-Etudes et de recherches en Statistiques et Développement 2 LERSTAD - laboratoire d-Etudes et de recherches en Statistiques et Développement 3 LERSTAD, Université Gaston Berger LERSTAD - laboratoire d-Etudes et de recherches en Statistiques et Développement 4 Institut National Polytechnique Yamoussoukro

Abstract : This paper considers estimation of the parameters for fractionally integrated processes with infinite varianceinnovations introduced by Kokoszka and Taqqu (1995). This is a finite parameter model which exhibitslong-range dependence and large fluctuations (heavy-tailed distributions). We introduce two proceduresto estimate the di erencing parameter and ARMA coecients. The first one is the Conditional Sum ofSquares (CSS) method and the second is the Minimum Hellinger Distance (MHD) estimator. Monte Carloexperiments are used to evaluate the finite sample performance of the proposed estimators, and compare itto the Markov chains Monte Carlo (MCMC) Whittle approach developed by Ndongo et al. (2010).

Keywords : Markov chains Monte Carlo. CSS method MHD estimator Long memory Whittle estimates Stable distributions





Author: Mor Ndongo - Abdou Kâ Diongue - Aliou Diop - Ouagnina Hili -

Source: https://hal.archives-ouvertes.fr/



DOWNLOAD PDF




Related documents