生长曲线模型的几乎无偏广义Liu估计

Almost Unbiased Generalized Liu Estimates in Growth Curve Model

针对生长曲线模型中所存在的复共线性问题,利用几乎无偏的思想,提出生长曲线模型参数矩阵的一种新的有偏估计——几乎无偏广义Liu估计,研究了该估计的优良性等性质,并证明了在均方误差意义下,几乎无偏广义Liu估计优于广义Liu估计及最小二乘估计,最后通过数据验证了相关结论.

Proposes a new biased estimation for parameter matrix,namely almost unbiased general Liu estimation.Then discusses its optimal property and proved that almost unbiased general Liu estimation is superior to general unbiased Liu estimation and least square estimation in the sense of mean square error. At last,gives a numerical example to illustrate theoretical results.