回归系数的stein型广义主相关估计

Combining Stein and Generalized Main Correlation Estimation of Regression Coefficient

根据主相关估计的思想,针对广义线性模型,本文给出了一种新的回归系数的有偏估计——广义主相关估计,并将其与stein型压缩估计相结合,提出stein型广义主相关估计,在均方误差的意义下考虑其特性,对偏参数的范围予以讨论.

Based on the estimation of the main correlation estimation idea for generalized linear models, we give a new biased estimate, which we call generalized main correlation estimation. Further, combining it and stein-type estimation, we propose the stein-type generalized main correlation estimation. Under the criteria of mean square error, we discuss its characteristics, and consider the scope of the bias parameter.