摘要
利用密度权导出了有限相关样本的期望估计,论证了相关密度权是对称的、非对角型矩阵;相关期望估计具有抗差性,且具有自动抗差功能.在运算过程中,其解与迭代初始值无关,其迭代初始值为误差方程的常数项,并且能保持变权运算后各变量的相关系数不变.
The density weight to refer the expectation estimation on the limited relative samples is applied in the research.It is demonstrated that the relative density weight is a kind of symmetrical and no diagonal matrix. The robust and auto-robust function is inherent in the relative expectation estimation. In the course of calculating, the solution has nothing to do with the beginning value of the iteration, which is the constant term of the error equation, and the correlation coefficient of each variable won′t change after the variant weight operation.
出处
《桂林工学院学报》
2004年第3期334-336,共3页
Journal of Guilin University of Technology
基金
广西自然科学基金资助项目(9642010)
关键词
密度权
相关期望估计
相关密度权
density weight
expectation estimation
relative density weight
