透過您的圖書館登入
IP:3.17.150.89
  • 學位論文

廣義伽瑪分配逐步型一區間設限資料之貝氏分析

The Bayesian Approach to Progressive Type-I Interval Censoring Data under Generalized Gamma Distribution

指導教授 : 林余昭

摘要


在存活分析實驗中, 常常因為某些原因, 必須以間斷的方式觀察實驗變化, 無法觀察完整實驗而產生區間設限。比較常見的是逐步型一區間設限。 本文假設逐步型一區間設限資料服從廣義伽瑪分配(GG) 。之後我們對GG 分配及統計方法做介紹, 其中統計方法包括馬可夫鏈蒙地卡羅法(MCMC) 、Gibbs 抽樣法及M-H 演算法。 我們利用不同統計方法去模擬參數估計值, 發現都很接近目標值, 其中MCMC 法的結果是最接近的。同時使用實際資料Carbone et al. (1967) 對模型參數做估計。最後將參數估計值與最大概似估計法(MLE) 及EM 演算法的結果做比較。

並列摘要


In the survival analysis, the experimental changes may not be continuously observed all the time.Hence complete observations are sometimes not available.In practice, only censored interval data can be obtained. In this research, we assume data come from the generalized gamma distribution and they are collected in progressive type-I interval ensoring. We then apply Bayesian analysis via MCMC to do the statistical estimation. Simulation studies, along with the mean square errors of parameters of interest, are shown.Moreover, we analyze the real data set, Carbone et al.(1967), and compare the results with previously done MLE and EM methods.

參考文獻


[26] 許文志, (2010) , 逐步型I區間設限資料的貝氏分析", 中原大學應用數學系碩士論
Methods, 30:1921-1935.
two-parameter exponential distribution under progressive censoring", Journal of
Applied Statistics, 25:707-714.
[3] Balakrishnan, N. and Aggarwala, Rita (2000) , Progressive Censoring: Theory,

延伸閱讀