A Comparative Study on the Estimation of the Parameters of the Markovian Processes- II

Document Type : Original Article

Authors

King Saud University, Saudi Arabia

Abstract

Consider the stationary autoregressive process with order one and non-zero mean. The weight parameter is taken to be priori distributed as uniform, standard normal and normal with zero mean. The Bayes' estimates are obtained and compared with conditional estimator for different values for the mean of the process. The error variance estimates by different methods are investigated. Some properties of the estimates for both the weight parameter and the error variance are discussed. The work is mainly performed through simulation and numerical integration techniques.

Keywords