MISSING DATA, IMPUTATION, AND THE BOOTSTRAP
MISSING DATA, IMPUTATION, AND THE BOOTSTRAP
复制标题
DOI：
10.2307/2290846
复制
发表时间：
1994-06-01
影响因子：
3.7
通讯作者：
EFRON, B
中科院分区：
文献类型：
--
作者：
EFRON, B
缺失数据是指由于缺少熟悉的数据结构的某些部分而变得困难的一类问题。例如，回归问题的预测向量中可能存在一些缺失值。本文涉及在缺失数据情况下评估估计器准确性的非参数方法。讨论了三个主要主题：缺失数据的引导方法、这些方法与多重插补理论的关系以及计算有效的执行方法。事实证明，最简单形式的非参数引导置信区间可以给出方便且准确的答案。引导方法和多重插补方法之间存在有趣的实践和理论差异，以及一些有用的相似之处。
Missing data refers to a class of problems made difficult by the absence of some portions of a familiar data structure. For example, a regression problem might have some missing values in the predictor vectors. This article concerns nonparametric approaches to assessing the accuracy of an estimator in a missing data situation. Three main topics are discussed: bootstrap methods for missing data, these methods' relationship to the theory of multiple imputation, and computationally efficient ways of executing them. The simplest form of nonparametric bootstrap confidence interval turns out to give convenient and accurate answers. There are interesting practical and theoretical differences between bootstrap methods and the multiple imputation approach, as well as some useful similarities.