A nonparametric likelihood ratio test
A nonparametric likelihood ratio test
复制标题
DOI：
10.1080/00949658308810702
复制
发表时间：
1983
影响因子：
1.2
通讯作者：
Richard L. Dykstara;R. Madsen;K. Fairbanks
中科院分区：
文献类型：
--
作者：
Richard L. Dykstara;R. Madsen;K. Fairbanks
利用似然比原理建立了一种非参数检验方法，用于检验随机排序是否为零假设。该检验也适用于检验单侧随机排序方案的分布是否相等。功效研究表明，在后一种情况下，该检验与Kolmogorov-Smirnov和Mann-Whitney-Wilcoxon相比是有利的，即使它提供了一些保护，以防止非随机排序的替代品。提供了小于或等于20的样本的百分比点
The likelihood ratio principle is employed to develop a nonparametric test for testing stochastic ordering as a null hypothesis. This test is also adapted for testing equality of distributions against one-sided stochastic ordering alternatives. Power studies indicate this test compares favorably with the Kolmogorov-Smirnov and Mann-Whitney-Wilcoxon in the latter situation, even though it provides some protection against alternatives that are not stochastically ordered. Percentage points for samples less than or equal to twenty are provided