Abstract
This paper presents a Bayesian method for identifying of moving average (MA) models. The proposed Bayesian methodology is based on replacing lagged errors of the original moving average model with appropriately lagged residuals from a long autoregressive. Then unlike Broemeling and Shaarawy [4], the exact structure of the approximation error when replacing true errors with corresponding residuals is derived and used in deriving the posterior probability mass function of the model order. The direct Bayesian technique is employed to identify the order of moving average models. The proposed and Broemeling and Shaarawy methods are compared using several simulation studies and a real data set. Simulation and a real data results show that the proposed method is superior to Broemeling and Shaarawy method in identifying the orders of moving average models.