本文主要目的在於探討誤差項具自我迴歸之貝氏線性混合效應模型。經由混合先驗分佈，找出顯著的固定效應與隨機效應變數，進而得到中位機率模型，並由該模型進行貝氏之推論與預測。最後將此模型應用在一組高血糖和相對高胰島素血症，對於葡萄糖的容忍度實驗之資料中，結果顯示，本文提出的方法可以提供準確的預測。 In this thesis, we address the problem of Bayesian linear mixed effects model with auto-regressive errors. We consider a mixture prior to identify subsets of covariates having nonzero fixed effect coefficients or nonzero random effects variance, and eventually obtain a median probability model, which is utilized for Bayesian inference and prediction. Finally, the proposed method is applied to a study of the association of hyperglycemia and relative hyperinsulinemia, and it yields very accurate prediction results.
