| 基于最大特征根的相对效率的下界估计 |
| |
| 引用本文： | 石爱菊.基于最大特征根的相对效率的下界估计[J].南京邮电大学学报(自然科学版),2008,28(3). |
| |
| 作者姓名： | 石爱菊 |
| |
| 作者单位： | 南京邮电大学数理学院,江苏南京,210003 |
| |
| 摘 要： | 定义了广义线性模型中参数的最小二乘估计(LSE)与最佳线性无偏估计(BLUE)的一种新的相对效率,这种相对效率定义为参数估计量的协方差阵的最大相对特征根之比.给出了当设计矩阵满秩,协方差阵非负定时这种新的相对效率的下界,研究了新的相对效率与广义相关系数之间的关系.由本文的讨论可见,当广义线性模型与主成分分析、典型相关分析方法结合应用时,这种相对效率具有很强的实用价值.
|
| 关 键 词： | 广义线性模型 相对特征根 相对效率 广义相关系数 |
A Lower Bound of the Relative Efficiency Based on the Largest Relative Eigenvalue |
| |
| SHI Ai-ju.A Lower Bound of the Relative Efficiency Based on the Largest Relative Eigenvalue[J].Journal of Nanjing University of Posts and Telecommunications,2008,28(3). |
| |
| Authors: | SHI Ai-ju |
| |
| Institution: | SHI Ai-ju College of Mathematics , Physics,Nanjing University of Posts , Telecommunications,Nanjing 210003,China |
| |
| Abstract: | This paper defines a new relative efficiency of the parameters of LSE with BLUE in singular G-M model.The new relative efficiency is based on the ratio of the largest relative eigenvalues of the covariance matrix.The paper gives a lower bound and a relation with the generalized relative coefficient where the design matrix is nonsingular in the model.The discussion,shows that the new relative efficiency is very useful in case where we combine general linear model with principal analysis and canonical correla... |
| |
| Keywords: | general linear model relative eigenvalues relative efficiency generalized relative coefficient |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
