Rank Based Inference
Rank Based Inference
复制标题
基于排名的推理
DOI：
--
复制
发表时间：
2005
期刊：
影响因子：
--
通讯作者：
S. Sheather
中科院分区：
文献类型：
--
作者：
T. Hettmansperger;J. McKean;S. Sheather
排名的使用有很长的历史，可以追溯到1945年威尔科松的开创性工作。早期的非参数等级检验是为基本的单样本和双样本设计开发的，然后扩展到单向和双向布局。这些检验因其非参数性质以及良好的效率和稳健性而广受欢迎。在这篇文章中，我们描述和说明如何将基于秩次的方法扩展到一般的线性模型，包括回归、方差分析和协方差分析。我们还描述了一个决定系数，它在评估模型与数据的适合性时很有用。
关键词：
等级检验；
R-估计数；
线性模型；
回归；
方差分析；
协方差分析
The use of ranks has a long history stretching back to 1945 with the seminal work of Wilcoxon. Early nonparametric rank tests were developed for the basic one- and two-sample designs and then extended to the one- and two-way layouts. These tests are popular because of their nonparametric property along with their good efficiency and robust properties. In this entry, we describe and illustrate how to extend rank-based methods to the general linear model including regression, analysis of variance, and analysis of covariance. We also describe a coefficient of determination that is useful when assessing the fit of a model to the data.
Keywords:
rank test;
R-estimates;
linear model;
regression;
analysis of variance;
analysis of covariance