离散型广义非线性模型包括Poisson,二项,负二项模型.本文讨论离敞型广义非线性纵向数据模型中偏离名义离差的检验问题,得到了检验的score统计量,并利用Monte Carlo方法研究了检验统计靖的性质.最后,利用杀虫剂数据说明了检验方法的应用.
Discrete exponential family nonlinear models include Poisson, binomial, and negative binomial models. This paper discusses the test for departures from nominal dispersion in the framework of discrete generalized nonlinear models with random effects for longitudinal data. The score test statistic is constructed, The properties of test statistic are investigated through Monte Carlo simulations. We illustrate our test method by the insecticide data.