Comparison of estimation methods for the density of autoregressive parameter in aggregated AR(1) processes
Articles
Dmitrij Celov
Vilnius University
Remigijus Leipus
Vilnius University, Institute of Mathematics and Informatics
Virmantas Kvedaras
Vilnius University
Published 2021-06-15
https://doi.org/10.15388/LMR.2007.24253
PDF

Keywords

random coefficientAR(1)
aggregation
disaggregation
long memory
mixture density

How to Cite

Celov , D. , Leipus, R. and Kvedaras, V. (2021) “Comparison of estimation methods for the density of autoregressive parameter in aggregated AR(1) processes”, Lietuvos matematikos rinkinys, 47(spec.), pp. 508–516. doi:10.15388/LMR.2007.24253.

Abstract

The article investigates the properties of two alternative disaggregation methods. First one, proposed in Chong (2006), is based on the assumption of polynomial autoregressive parameter density. Second one, proposed in Leipus et al. (2006), uses the approximation of the density by the means of Gegenbauer polynomials. Examining results of Monte-Carlo simulations it is shown that none of the methods was found to outperform another. Chong’s method is narrowed by the class of polynomial densities, and the secondmethod is not effective in the presence of common innovations.Bothmethodswork correctly under assumptions proposed in the corresponding articles.

PDF

Downloads

Download data is not yet available.