摘要
针对多元线性模型未知参数最小二乘估计的相对效率难以精确计算的问题，分别利用矩阵行列式和矩阵的迹定义了两种新的相对效率ｅ０（Ｂ）和ｅ１（Ｂ）。并且在不同的模型条件下，分别给出了ｅ０（Ｂ）和ｅ１（Ｂ）不同的下界估计。讨论了这两种效率的相互关系以及它们与广义相关系数的关系，这对于线性模型中最小二乘估计的度量是有益的。
Two relative efficiencies of least square estimates for unknown parameters in multiple linear models are defined. Because it is difficult to calculate the relative efficiencies exactly,different subbounded estimates in different models are given.Meanwhile, relationships between the two relative efficiencies and between them and general relative coefficient are established. These results are helpful to appraise least square estimates in linear models.
出处
《江汉石油学院学报》
CSCD
北大核心
1996年第2期119-123,共5页
Journal of Jianghan Petroleum Institute
基金
江汉石油学院科研发展基金
关键词
多元回归分析
线性模型
最小二乘法
参数估计
multivariate regression analysis
linear models
least squares methods
parameter estimations
unbiased estimations
(relative effciency)
