具冪次轉換，隨機效應及ARMA(p,q)相依之成長曲線模式的貝氏分析

Abstract

本論文是討論具冪次轉換，隨機效應與ARMA(p,q)相依之一般成長曲線模式的貝氏分析．我們提出二種不同的事前分配，並利用有用的逼近法及 MCMC法．而且在參數估計及未來值的預測問題作一比較．

In this paper we devoted ourselves to a general growth model with power transformation, random effects and ARMA(p,q) dependence. Two prior distributions are proposed and put into comparisions in parameter estimation and predition of future values. Useful approximations and MCMC methods are employed. Some numerical results are illustrated with real and simulated data.