一阶半变系数混料模型的D-最优设计  

D-Optimal Design for First-order Semi-Varying Coefficient Mixture Model

作　　者：刘妙玲 张崇岐[1] LIU Miao-ling;ZHANG Chong-qi(School of Mathematic and Science Information,Guangzhou University,Guangdong Guangzhou 510006,China)

机构地区：[1]广州大学数学与信息科学学院

出　　处：《数理统计与管理》2018年第6期1041-1049,共9页Journal of Applied Statistics and Management

基　　金：国家自然科学基金(11671104);广州大学全日制研究生“基础创新”项目(2017GDJC-M31)

摘　　要：通过研究一阶半变系数混料回归模型,转变成混合效应混料回归模型,进而得到模型的信息矩阵。针对一阶半变系数混料回归模型,求出其单纯形-中心设计的D-最优设计和G-最优设计,由此得到一阶半变系数混料回归模型中D-最优设计与G-最优设计不等价。By studying first-order mixture regression model with semi-varying coefficient, it is obtained the mix-effect mixture regression model by using the variant method and then obtained the information matrix. The D-optimum design and G-optimum design for first order mixture regression model with semi-varying coefficient are calculated. Meanwhile, it is showed that there are not equivalent between D-optimum design and G-optimum design.

关 键 词：半变系数 混料回归模型 信息矩阵 最优设计 

分 类 号：O212[理学—概率论与数理统计]