現狀資料下利用經驗概似比值法探討雙樣本之檢定;Empirical Likelihood Ratio Tests for Two Sample Comparison under Current Status Data

National Chung Cheng University Institutional Repository > 理學院 > 數學系統計科學研究所 > 學位論文 > Item A095B0000Q/316

Please use this identifier to cite or link to this item: http://ccur.lib.ccu.edu.tw/handle/A095B0000Q/316

Title:	現狀資料下利用經驗概似比值法探討雙樣本之檢定;Empirical Likelihood Ratio Tests for Two Sample Comparison under Current Status Data
Authors:	侯科宇;Hou, Ke-Yu
Contributors:	數學系統計科學研究所
Keywords:	現狀資料;經驗概似比值;雙樣本檢定;Current status data;Empirical likelihood ratio;Two sample test
Date:	2017
Issue Date:	2019-07-17
Publisher:	數學系統計科學研究所
Abstract:	本篇論文主要是探討現狀資料下的雙樣本檢定。基於 Banerjee & Wellner (2001)，我們建構限制下與未限制下存活函數的最大概似估計。之後建構雙樣本經驗概似比值檢定統計量來比較雙樣本存活曲線。我們應用拔靴法的方式建構虛無假設下的分配並且計算 p-value。利用模擬試驗檢驗我們提出來的方法，再與另一個檢定雙樣本現狀資料的方法 Groeneboom (2012) 做比較。最後運用我們提出來的方法分析一筆真實資料。 This thesis focuses on the two sample test of the survival function of the failure time underthe current status data. Based on Banerjee & Wellner (2001), we construct the unconstrainedand the constrained maximum likelihood estimation of the survival function of the failure time.Then, we construct the empirical likelihood ratio test for two sample comparison. For the pvaluecomputation, we apply the bootstrap method to construct the null hypothesis distribution.We examine the finite-sample performance of the proposed approach by simulation studies andcompare it with the method by Groeneboom (2012) which is also a two sample test for currentstatus data. Finally, we also apply our proposed methodology to analyze a real data example.
Appears in Collections:	[數學系統計科學研究所] 學位論文

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	745	View/Open

Loading...