摘要
在未知线性回归模型测量误差的协方差矩阵和非零均值(由非模型化误差及测量设备系统性误差组成)情况下,提出一种回归参数的“最小均方误差估计“,证明了其优良性。还对稳健估计、测量精度分析中随机与系统性误差分离、白塞尔公式与变量差分法统计随机方差结果不一致等问题,作了深入研究分析。
Under the condition of unknown covariance matrix and nonzero mean valu composed of non-model error and systematic error of measuring equipment) of measuring error in a linear regression model, a “minimum mean squre error estimation” is presented and proven to be superior. In addition, a thorough study and analysis is done to the problems of robust estimation, separation of systematic error and random error in measuring accuracy analysis, inconsistency of random variance statistics through variable difference with that through Bessell formula, etc.,
出处
《宇航学报》
EI
CAS
CSCD
北大核心
1991年第4期19-25,共7页
Journal of Astronautics
关键词
估计理论
数学模型
参数
矩阵
Estimation theory, mathematical models, Parameter estimation, Mean square error matrix estimation, Robust estimation, Question analysis.