摘要
在一类非线性模型中，本文利用了文［３］投影似然方法，推导出了拟似然得分函数是线性无偏估计函数类中唯一最优估计函数，推广了古典线性模型中著名的Ｇａｕｓｓ－Ｍａｒｋｏｖ定理。最后，证明了拟似然估计（ＱＬＥ）比最大似然估计（ＭＬＥ）效率低，从而进一步证实了文［２］的结论。
This paper,in nonlinear models,the quasiscore funtion is derived as a projection of the unknow true score function based method of the projected likelihood presented by Ncleish and Smaal(1992).The uniquess and opitimality fo quasiscore function as estimating function is proveed in the class of linear unbaise estimating function.Furthmoore,it is proved that the quasi likelihood estimation is less efficient than Maximm likelihood estimation, the result of McCllagh and Nelder(1989) is demonstrated.
出处
《工程数学学报》
CSCD
1996年第2期59-64,共6页
Chinese Journal of Engineering Mathematics
