广义Gamma分布簇广义线性混合模型的构建

 本文在广义Gamma分布簇基础上引入异质性来构建广义线性混合模型。本文构建的广义Gamma分布簇广义线性混合模型在广义线性混合模型的框架下分析，通过参数重整技术把广义Gamma分布簇变量的建模问题与指数分布簇变量的建模问题联系起来，模型推断可以方便地利用广义线性混合模型和广义线性模型的研究成果，同时也可以方便地推广到其他模型。三参数广义Gamma分布可以收缩到两参数的Gamma分布、Weibull分布或指数分布，能降低模型误设的风险，还能便利地分析误差结构。

Generalized Linear Mixed Models based on Generalized Gamma Distribution

Generalized Gamma Distribution Family models can be embedded into Generalized Linear Mixed Models by way of special link function. Models Estimation can be easily achieved with the theory support of Generalized Linear Mixed Models. Some other models such as random intercept models, variable-coefficient models, repeated models, spatial models and so on can easily be derived from this Generalized Linear Mixed Models based on Generalized Gamma Distribution. The aim of constructing is to reduce model misspecification risk as well as analyze error structure.