Introduction to the Special Issue: Nonparametric Statistics
Introduction to the Special Issue: Nonparametric Statistics
复制标题
特刊简介：非参数统计
DOI：
10.1214/088342304000000765
复制
发表时间：
2004
影响因子：
5.7
通讯作者：
G. Casella
中科院分区：
文献类型：
--
作者：
R. Randles;T. Hettmansperger;G. Casella
近年来，非辅助医学领域持续快速增长。这在一定程度上可能是由于计算能力的互补增加，但可能更多地反映了需要更灵活和更复杂的模型来描述我们面临的不断增加的数据量和复杂性。这种史无前例的增长表明，需要一期关于非参数统计的特刊，而《客座编辑》负责编写这样一期，作为进入这一浩瀚知识体系的入口。众所周知，当代非参数统计所包含的内容远远超过传统的无分布秩检验法及其相应的R-估计方法，这些方法是为简单分析方差设计而发展起来的。事实上，非参数统计可以而且应该被广泛地定义为包括所有不使用基于单一参数族的模型的方法。现在被归入非参数方法的领域包括一般线性模型(包括多变量数据结构、非参数生存分析、非参数曲线估计和Bootstrap方法)以及本期文章中所说明的其他领域。传统的方法在许多好的文本中都有很好的记载，包括Wilcoxon符号秩和检验和Hodges-Lehmann估计，Wilcoxon-Mann-Whitney秩和检验和Kruskal-Wallis检验。其中一些方法可以追溯到20世纪40年代和50年代，它们的流行在很大程度上要归功于埃里希·莱曼和其他人的工作
In recent years, the field of nonparametrics has continued to grow at a rapid rate. This may be partially due to the complementary increase in computing power, but is perhaps more reflective of the need for more flexible and complex models to describe the everincreasing amount and complexity of data that face us. This unprecedented growth has signaled the need for a special issue on nonparametric statistics, and the Guest Editors were charged with compiling such an issue, one that would serve as an entrance to this vast body of knowledge. As we know, contemporary nonparametric statistics embraces far more than traditional distributionfree rank tests and their corresponding R-estimation methods developed for simple analysis of variance designs. Indeed, nonparametric statistics can and should be broadly defined to include all methodology that does not use a model based on a single parametric family. Now included under the rubric of nonparametric methods are such diverse fields as general linear models (including multivariate data structures, nonparametric survival analysis, nonparametric curve estimation and bootstrap methods), as well as others illustrated in the articles in this issue. The traditional methods are well documented in many good texts and include the Wilcoxon signed rank test and Hodges–Lehmann estimate, the Wilcoxon–Mann–Whitney rank sum test and the Kruskal–Wallis test. Some of these methods date back to the 1940s and 1950s, and much of their popularity is due to the work of Erich Lehmann and others