线性回归的几乎无偏两参数岭M估计
On almost unbiased two-parameter ridge-type M-estimatorfor linear regression model
胡宏昌 1闾欣萍1
作者信息
- 1. 湖北师范大学 数学与统计学院,湖北 黄石 435002
- 折叠
摘要
针对线性回归模型中设计矩阵存在复共线性问题,利用几乎无偏估计和两参数岭估计的思想,提出了几乎无偏两参数岭M估计.在均方误差准则下,得到了几乎无偏两参数岭M估计优于两参数岭M估计、几乎无偏两参数岭估计,并通过数值模拟验证了估计方法和结论的有效性.
Abstract
Considering the existence of multicollinearity in the design matrix of the linear model,this paper proposed a new ridge-type M-estimatorfor almost unbiased two-parameter based on almost unbiased estimation and two parameter ridge-type M-estimator.Under the mean square errorcriterion,the proposed estimator outperforms two-parameter ridge M-estimator and the almost unbiased two-parameter ridge estimate.Moreover,a numerical example experiment was presented to verify the validity of the estimation method and conclusions.
关键词
线性回归模型/均方误差/两参数岭M估计/几乎无偏两参数岭M估计Key words
linear regression model/mean square error/two-parameter ridge M-estimate/almost unbiased two-parameter ridge-type M-estimate引用本文复制引用
出版年
2024
NETL
NSTL
万方数据