Box-Cox轉換之線性模型上Sweeting的先驗分配

Abstract

利用貝氏法來分析觀察值經由Box-Cox轉換的線性模型，文獻上有些不同的先驗分配。本論文將比較這些先驗分配 (含MCMC法應用於Sweeting之先驗分配) 的優劣點，並給真實資料及模擬的結果。

For the analysis of observations in the linear model with Box-Cox transformation, via Bayesian approaches, some different priors have been proposed. In this thesis, we will compare relative merits of different priors including the prior proposed by Sweeting using MCMC methods. Some numerical results with a set of real data and simulation are given.