Wan J Y, Galecki A T
Department of Biostatistics, University of Michigan, Ann Arbor.
Methods Inf Med. 1992 Sep;31(3):215-8.
A Poisson regression model is proposed for the analysis of incidence rates presented in a two-way table classified by two categorical variables. It is shown that the likelihood function is the same as that using Glasser's exponential covariate model. An algorithm is given to solve the maximum likelihood estimates of the regression parameters. The model is evaluated via deviance and the method is illustrated with an example. Some extensions of the model are discussed.
提出了一种泊松回归模型，用于分析由两个分类变量分类的双向表中呈现的发病率。结果表明，似然函数与使用格拉斯指数协变量模型时的似然函数相同。给出了一种求解回归参数最大似然估计的算法。通过偏差对模型进行评估，并通过一个例子对该方法进行说明。讨论了该模型的一些扩展。
