Multiple comparison procedures based on robust statistics

在多样本模型中，我们考虑了位置参数的多重比较。提出了基于统计量最大值的一步多重比较法，并建立了该方法的理论基础。其次，在所提出的单步多重比较检验的基础上，推导出封闭检验程序。这些拟议的程序完全是保守的。进一步讨论了多重比较的渐近理论。特别地，在渐近理论中，序列拒绝过程可以在不等样本量下构造。该方法优于其他多重比较法。

We considered multiple comparisons for location parameters in multi-sample models. The single-step multiple comparison procedures based on the maximum values of statistics are proposed and the theory of the proposed procedures constructed. Next closed testing procedures are derived based on the proposed single-step multiple comparison tests. These proposed procedures are exactly conservative. Furthermore the asymptotic theory for the multiple comparisons is discussed. Especially sequentially rejective procedures can be constructed under unequal sample sizes in the asymptotic theory. The proposed procedures are superior to the other multiple comparison procedures.