一类线性混合模型的广义估计方程方法应用  

Application of Generalized Estimation Equation Methodfor a Class of Linear Mixed Models

作　　者：余俊飞 赵慧秀[1] YU Jun-fei;ZHAO Hui-xiu(School of Science,Nanjing University of Science and Technology,Nanjing 210094,China)

机构地区：[1]南京理工大学理学院,江苏南京210094

出　　处：《兰州文理学院学报（自然科学版）》2021年第4期1-4,共4页Journal of Lanzhou University of Arts and Science(Natural Sciences)

摘　　要：对于纵向数据下含有两个方差分量的线性混合模型,基于广义估计方程方法估计模型中的固定效应和方差分量.首先计算模型的边际均值和边际协方差矩阵,然后通过迭代估计方程求出模型中未知参数的估计.该方法不假设响应变量的分布,回归参数的标准差较小,能够得到稳健的估计结果.实例分析和随机模拟结果表明提出的方法是可行的.For the linear mixed model with two variance components under longitudinal data,the fixed effects and variance components in the model were estimated based on the generalized estimation equation method.First,the marginal mean and marginal covariance matrix of the model were calculated,then the unknown parameters in the model were estimated by iterative estimation equations.This method did not assume the distribution of response variables,and the standard deviation of regression parameter estimation was small,so that it can obtain robust parameter estimation results.Example analysis and random simulation results showed that the proposed method was feasible.

关 键 词：广义估计方程 线性混合模型 固定效应 方差分量 

分 类 号：O212.2[理学—概率论与数理统计]