The Study of Censored Data with Group Decision

碩士論文[[abstract]]基於貝氏分析法的精神，決策者主要以風險為考慮對象，常根據以往的經驗及資料，考慮適當的事前機率，再據以得到事後機率，使得決策者對整個可能發生的風險可以掌握，有關先前的諸多抽樣檢驗的研究中，決策者對於參數的事前分配各具有獨到的看法，雖然已經解決了許多問題，但是由於現今的公司當中，往往是由一個經營團隊來做決策，如果能整合公司裡各個專家的見解再做出決策，或許能夠更加改善許多決策結果。本研究乃在擴展Lin(2001)設限資料下指數母體的貝氏抽樣設計，在第二章中探討型 設限下的貝氏抽樣設計模型，在第三章中探討均勻設限資料下的貝氏抽樣設計模型，將其引入集體決策的觀念，在相同的權重下，探討共軛事前分配的整合對風險值的影響，並且對於相關的參數作敏感度分析，觀察各參數的變動對貝氏風險函數值的影響。[[abstract]]Based on the spirit of Bayesian Analysis, risk is the main concern of decision makers. They usually consider appropriate prior probability according to their past experience and data to get posterior probability and get hold of all probable risk. There are lots of studies about sampling inspection plans in literature and shoe original viewpoints towards the prior distribution of parameter. However, a team often makes a company’s decision. If we can integrate the opinions from many experts to make a decision, the decision can be more finely made.This thesis extends Bayesian sampling plans based on censored data of Lin (2001). In chapter 2, we confer Bayesian sampling plans under the Type censoring model. In chapter3, we confer Bayesian sampling plans under the Uniform random censoring. Furthermore, we introduce group decision into this thesis and discuss the influence of the integration of risk on the conjugate prior distribution with the weight being the same, Besides, we will make the sensitive analysis for relational parameters and observe the effect of Bayes risk values