交叉试验设计分析的拟线性模型

A OUASILINEAR MODEL FOR ANALYSING DATA OF CROSSOVER TRIALS

本文针对输出为计量资料的情形,利用适用面较广的拟线性模型,揭示输出与输入史之间关系。其中特别考虑到输入的累积残余效应。利用极大似然估计和似然比检验的方法对参数进行估计和检验。这些都在IBM-PC/XT机上以FORTRAN 77实施。就一次给药情形对本模型进行了模拟研究,从而考察参数估计方法的优良性。初步模拟结果表明,参数估计精度满意。本模型不需设置任何预检验,无论有否残余效应都可应用。本方法还可获得现存方法不能获得的结果。

A quasilinear model considering output, history of input and cumulative residual effect isproposed for analysing measurement data of crossover trials. The maximum likelihood criterion andlikelihood ratio test are used to estimate and test parameters in the model. This is performed on theIBM/XT with FORTRAN 77 programs. Computer simulation shows that the model is effective. Thismodel aan be used in any case and yields some results which can not be got with other modelsdescribed in literature.