Maximum Likelihood Estimation for Short Time Series with Replicated Observations: A Simulation Study

by Shelton Peiris,R.Mellor and P.Ainkaran.

Abstract: Analysis of a large number of independent replications from short, first order autoregressive type time series is considered. Maximum likelihood estimation (mle) procedure is discussed in both approximate and exact forms. A simulation study is carried out. It is shown that both the approximate and exact mle methods provide unbiased and very efficient (in the minimum mean square sense) estimates for the parameters.

Key Words: Time series, Correlation, Estimation, Bias, Repeated measurements, Efficient, Maximum likelihood, Minimum mean square

Authors:
Shelton Peiries, shelton@maths.usyd.edu.au
R. Mellor, r.mellor@uws.edu.au
>P. Ainkaran, karan@maths.usyd.edu.au

Editor: Hatemi-J, Abdulnasser,Abdulnasser.Hatemi-J@ish.his.se

READING THE ARTICLE: You can read the article in portable document (.pdf) format (109756 bytes.)

NOTE: The content of this article is the intellectual property of the authors, who retains all rights to future publication.

This page has been accessed 369 times since July 24, 2006.

Return to the InterStat Home Page.