线性模型中鞅差误差条件下回归系数L.S.估计的平均相合性

CONSISTENECY IN R-TH MEAN OF LEAST SQUARES ESTIMATES IN LINEAR REGRESSION MODELS WITH MARTINGALE DIFFERENCE ERRORS

本文中,我们讨论了线性回归模型y_i=X_i^(?)β+e_i i=1,2,…的随机误差(e_i)是鞅差序列的条件下,给出了回归系数β的最小二乘估计的r-平均相合性.

Suppose linear regression models: yi=xi β+ei i= l,2, …?where β is a unknown regression coefficient vector. βn is βL.S. estimate about observed points:y1, y2,… yn·βn is said to be r-th mean consistency estimate , if there is a constant r>0, such thatIn this paper, the authors have discussed that consistency in r- th mean of least squares estimates in linear regression models under the condition that random error {ei} is a sequence of martingale difference.