Maximum-Likelihood Estimation in a Special Integer Autoregressive Model

Jung, Robert C and Tremayne, Andrew R (2020) Maximum-Likelihood Estimation in a Special Integer Autoregressive Model. Econometrics, 8 (2). p. 24.

Access the full-text of this item by clicking on the Open Access link.

Text
econometrics-08-00024-v2.pdf - Published version
Download (679kB) | Preview

Official URL: https://www.mdpi.com/2225-1146/8/2/24

Abstract

The paper is concerned with estimation and application of a special stationary integer autoregressive model where multiple binomial thinnings are not independent of one another. Parameter estimation in such models has hitherto been accomplished using method of moments, or nonlinear least squares, but not maximum likelihood. We obtain the conditional distribution needed to implement maximum likelihood. The sampling performance of the new estimator is compared to extant ones by reporting the results of some simulation experiments. An application to a stock-type data set of financial counts is provided and the conditional distribution is used to compare two competing models and in forecasting.

Item Type:	Article
Uncontrolled Keywords:	autoregression, counts, maximum-likelihood, binomial-thinning
Depositing User:	Symplectic Admin
Date Deposited:	09 Sep 2020 09:55
Last Modified:	17 Mar 2024 08:43
DOI:	10.3390/econometrics8020024
Open Access URL:	http://doi.org/10.3390/econometrics8020024
Related URLs:	Author Publisher
URI:	https://livrepository.liverpool.ac.uk/id/eprint/3100516