![]() | Only 14 pages are availabe for public view |
Abstract The main contribution of study is using the bayesian generalized least square approach and Ismail’s modification to broemeling and Shaarawy’s method to derive the approximate predictive densities for autoregressive moving average models. the comparisions between the proposed approximations and Newbold, Zellner and Reynolds and Broemeling and Shaarawy approximation are done via several simulation studies using different ARMA models. A Gibbs sampling algorithm to compute predictive density of ARMA models in developed. then comparing the results with that resulting from analytic approximations. using simulation to check the efficiency of the proposed Gibbs sampling algorithm. ten real time series are analyzed. |