A New Method of Estimation in a Moving Average Model of Order One.
Abstract
The exact likelihood of the data coming from a moving average model of order is complicated. In this paper, the authors propose a method of estimation of the parameters of a moving average model of order one based on the approximate likelihood of the data and on the simulation of a pair of random variables. Some comparisons were made of this method with some well known methods for moderate sample sizes. A computer program is appended which is helpful in using this methods. Keywords: Time series; Analysis; Computerized simulation; Multivariate analysis.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1986
- Accession Number
- ADA186039
Entities
People
- H. A. Chapeh
- M. B. Rao
Organizations
- University of Pittsburgh