右設限和現狀數據混合資料下累積風險之研究

Cumulative hazard rate estimation for mixed right-censored and current status data

林智謙(Chin-Chien Lin)

指導教授 ： 吳裕振

中原大學/理學院/應用數學研究所/碩士(2015年)

https://doi.org/10.6840/cycu201500152

摘要

存活的統計分析，在醫學與電子學上是非常重要的。本篇論文研究右設限資料和現狀數據混合對累積風險函數去做最大概似估計。在資料型態一樣下，陳奕峻(2015)的初稿中，他是用伯式多項式去描述累積風險函數，並用貝氏方法去估計。跟本篇論文的不同是，我們是用最大概似估計。因為參數複雜，並不好計算，我們將利用馬可夫鏈蒙地卡羅─Simulated Annealing來計算，並且做200個資料和50個資料做模擬比較，其結果符合我們的理論，我認為此方法可避開一些繁雜的計算，所以可提供給學者和專家參考與分析。

關鍵字

累積風險函數 ； 最大概似估計 ； 現狀數據 ； 右設限

並列摘要

The statistical analysis of survival is very important in medicine and electronics. The study mixed the right-censored and current status data to maximum likelihood estimate the cumulative hazard function. Based on the same data type, the first script wrote by Yi-Chung Chen (2015) was estimated by means of Bernstein polynomial and Bayesian distribution function.Contrary to this research, the researchers got the results by the maximum likelihood estimator method. We will use the Markov chain Monte Carlo method─Simulated Annealing to calculate the parameters bacause they are too complicated to calculate and the researchers compared 200 documents to 50 documents, whose results consisted with the researcher thinksthat with the approach, some complicated calculation can be avoided and offered to professors and specialists as a reference.

並列關鍵字

cumulative hazard function ； maximum likelihood estimate ； current status data ； right-censored

參考文獻

[7 ] 許時淮 (2015) 現狀數據資料下勝算比之研究，初稿

[2 ]G. Casella , R.L. Berger ” Statistical inference ” , Duxbury Press 1990.

[3 ]Christian P. Robert, George Casella ” Monte Carlo Statistical Methods

[4 ]P. Green ” Reversible jump Markov chain Monte Carlo computation and

Bayesian model determination ”