具有正交约束的多因子可加性实验模型参数估值的无偏性

    Unbiasedness of Parameter Estimates In The Model of Additional Experiment Under The Condition of Orthogonal Constranits

    • 摘要: 一个好的实验模型选择的主要准则之一,是模型中参数的估值应该是无偏的。本文在"多因子可加性实验模型参救估值及唯一性"一文的基础上,进一步证明了在因子间各级交互效应满足正交约束及正态误差分布条件下,模型参数估值是无偏的。

       

      Abstract: A fundamental criterion for selecting a proper model in speci fied experiment requires the unbiased estimates of parameters in the model. In terms of author's former paper "Estimate of Parameters and its Uniqueness in a Model of Additional Experiment with Multifactors" a rigorous description is given in this work, in which the parameter estimates in the model are unbiased under the condition of the orthogonal imposed on all interactions among the factors, and the condition of normal distribution of the errors.

       

    /

    返回文章
    返回