摘要
讨论了半相依回归系统(SURS)中回归系数的广义协方差改进估计(GCIE)的效率问题,得出了几种估计的效率度量公式在极限意义下等价的结论.以该结论为基础,我们对一些简单的协方差结构的SURS中回归系数的GCIE系列进行计算机模拟,得到了m=3时,在这些模型下,GCIE的极限便是最佳线性无偏估计的结论.
In this paper,the effeciency of the generalized covariance improvement estimate (GCIE) of the regression parameters (RP) in seemingly unrelated regression system is discussed.The result that several measument formulas of the effeciency are equivelant in the sense of the limit is given.Based on the result,some models of SURS with simpler covariance construction are examined by means of computer simulation.Simulation results show that the limit of GCIE is the best linear unbiased estimate (BLUE) of the RP for the models.
出处
《江西师范大学学报(自然科学版)》
CAS
1999年第3期213-218,共6页
Journal of Jiangxi Normal University(Natural Science Edition)
基金
江西省自然科学基金
关键词
估计效率
半相依回归模型
广义协方差改进估计
estimation effeciency
seemingly unrelated regression system(SURS)
generalized covariance improvement estimate(GCIE)