Parameter estimation for generalized random coefficient autoregressive processes |
| |
Institution: | 1. Sookmyung Women''s University, Seoul, South Korea;2. Department of Statistics, The University of Georgia, 204 Statistics Building, Athens, GA 30602-1952, USA;1. University of Bern, Institute of Mathematical Statistics and Actuarial Science, Alpeneggstrasse 22, CH-3012 Bern, Switzerland;2. Department of Statistics, University of Washington, Seattle, WA 98195-4322, United States |
| |
Abstract: | A generalized random coefficient autoregressive (GRCA) process is introduced in which the random coefficients are permitted to be correlated with the error process. The ordinary random coefficient autoregressive process, the Markovian bilinear model and its generalization, and the random coefficient exponential autoregressive process, among others, are seen to be special cases of the GRCA process. Conditional least squares, and weighted least-squares estimators of the mean of the random coefficient vector are derived and their limit distributions are studied. Estimators of the variance-covariance parameters are also discussed. A simulation study is presented which shows that the weighted least-squares estimator dominates the unweighted least-squares estimator. |
| |
Keywords: | |
本文献已被 ScienceDirect 等数据库收录! |
|