Nonparametric estimators of survival function with left-truncate current status data

碩士 === 東海大學 === 統計學系 === 101 === Current status data arise from the basic version of the interval censoring where the observation consists only of an examination time and knowledge of whether the failure time has occurred before the examination time. In some cases, the failure time also suffers left...

Full description

Bibliographic Details
Main Authors: Chen, Mao-Cyuan, 陳茂銓
Other Authors: Shen, Pao-Sheng
Format: Others
Language:en_US
Published: 2013
Online Access:http://ndltd.ncl.edu.tw/handle/56575845312537809784
Description
Summary:碩士 === 東海大學 === 統計學系 === 101 === Current status data arise from the basic version of the interval censoring where the observation consists only of an examination time and knowledge of whether the failure time has occurred before the examination time. In some cases, the failure time also suffers lefttruncation, which results in left-truncated current status (LTCS) data. In this article, we first point out that for LTCS data the nonparametric estimator using Turnbull's EM algorithm is not the NPMLE since left-truncation times can also be left-censoring times. However, based on innermost sets, we can still obtain a nonparametric estimate using Turnbull's algorithm. Simulation study indicates that both estimators perform adequately and are consistent.