具有双重删失数据的非参数假设检验和置信区间

Non-parametric hypothesis testing and confidence intervals with doubly censored data.

作者信息

Chen Kun, Zhou Mai

机构信息

Department of Statistics, University of Kentucky, Lexington, KY 40506-0027, USA.

出版信息

Lifetime Data Anal. 2003 Mar;9(1):71-91. doi: 10.1023/a:1021834206327.

DOI:10.1023/a:1021834206327

PMID:12602775

Abstract

The non-parametric maximum likelihood estimator (NPMLE) of the distribution function with doubly censored data can be computed using the self-consistent algorithm (Tumbull, 1974). We extend the self-consistent algorithm to include a constraint on the NPMLE. We then show how to construct confidence intervals and test hypotheses based on the NPMLE via the empirical likelihood ratio. Finally, we present some numerical comparisons of the performance of the above method with another method that makes use of the influence functions.

摘要

具有双重删失数据的分布函数的非参数极大似然估计器（NPMLE）可使用自洽算法来计算（滕布尔，1974年）。我们将自洽算法进行扩展，使其包含对NPMLE的一个约束。然后我们展示如何通过经验似然比基于NPMLE构建置信区间并检验假设。最后，我们给出上述方法与另一种利用影响函数的方法在性能方面的一些数值比较。