• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文学位 >Empirical likelihood-based adjustment methods.
【24h】

Empirical likelihood-based adjustment methods.

机译：基于经验似然的调整方法。

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Auxiliary information is frequently used in survey sampling at the estimation stage to increase the precision of estimators. The class of calibration estimators introduced by Deville and Sarndal (1992) is obtained by replacing the design weights with the so-called calibration weights, i.e. the closest weights from the design weights (with respect to a given divergence measure) that satisfy the constraints that incorporate the auxiliary information.; The presence of covariate information in experimental design situations is similar to the presence of auxiliary information in survey sampling. Methods for nonparametric covariance adjustment from Quade (1967), Puri and Sen (1971), Koch et al. (1982), and Koch et al. (1998) are similar to the calibration methods for the two survey samples described by Zieschang (1990). In all these cases the covariance-adjusted estimators are obtained from the unadjusted estimators by minimizing a quadratic criterion subject to equal means constraints on the covariates.; Ideas from survey sampling are used to generalize previously mentioned methods for nonparametric covariance adjustment. First, an empirical likelihood-based adjustment method is proposed for the construction of confidence intervals for the difference between means. A stratified version of the method and related methods that use criteria based on other divergence measures are described. Next, empirical likelihood-based adjustment methods are developed for the difference between more general parameters of interest under more general constraints. Finally, alternative empirical likelihood-based methods, that use a weighted empirical likelihood criterion, are developed for the construction of confidence intervals for the difference between means and stratified versions.
机译：在估计阶段，辅助信息经常用于调查抽样中，以提高估计器的精度。 Deville和Sarndal（1992）引入的校准估计器类别是通过将设计权重替换为所谓的校准权重来获得的，即，校准权重与设计权重（相对于给定的发散度量）最接近，可以满足以下约束条件：合并辅助信息。实验设计情况下协变量信息的存在类似于调查抽样中辅助信息的存在。非参数协方差调整的方法来自Quade（1967），Puri和Sen（1971），Koch等。 （1982）和Koch等。 （1998年）类似于Zieschang（1990年）描述的两个调查样本的校准方法。在所有这些情况下，通过最小化二次标准，对协变量进行均等约束，可以从未调整的估计量中获得协方差调整的估计量。来自调查抽样的想法被用来概括前面提到的非参数协方差调整方法。首先，提出了一种基于经验似然性的调整方法，用于构造均值之间差异的置信区间。描述了使用基于其他差异度量的标准的方法和相关方法的分层版本。接下来，针对在更一般的约束条件下感兴趣的更一般的参数之间的差异，开发了基于经验似然的调整方法。最后，开发了使用加权经验似然准则的基于经验似然的替代方法，以构建均值和分层版本之间差异的置信区间。

著录项

  • 作者

    Luta, Gheorghe.;

  • 作者单位

    The University of North Carolina at Chapel Hill.;

  • 授予单位 The University of North Carolina at Chapel Hill.;
  • 学科 Biology Biostatistics.; Statistics.
  • 学位 Ph.D.
  • 年度 2006
  • 页码 157 p.
  • 总页数 157
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类 生物数学方法;统计学;
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号