随机约束条件下一种新的加权混合两参数估计  

A New Weighted Mixed Two-parameter Estimator Under Stochastic Constraints

作　　者：黄伯强[1] Huang Boqiang(Nanjing Normal University Zhonghei College,Zhenjiang Jiangsu 212334,China)

机构地区：[1]南京师范大学中北学院,江苏镇江212334

出　　处：《统计与决策》2021年第23期16-20,共5页Statistics & Decision

基　　金：国家社会科学基金资助项目(18BTJ040)。

摘　　要：为解决线性模型的多重共线性问题,文章通过结合几乎无偏Liu算子T_(d)与岭算子W_(k),提出了随机约束条件下一种新的加权混合两参数估计β_(MWMTPE),并在均方误差准则下,把新估计与已有的加权混合估计β_(WME)、加权混合岭估计β_(WMRE)、加权混合Liu估计β_(WMLE)、加权混合几乎无偏Liu估计β_(WAULE)进行了比较,给出了新估计优于其他估计的充分必要条件。最后,通过具体数据实例对理论结果进行了验证,结果表明在合适的参数取值条件下,新估计优于其他几个加权混合估计。In order to solve the multicollinearity of linear models,this paper proposes a new weighted mixed two-parameter estimator β_(NWMTPE) by combining almost unbiased Liu operator T_(d) and ridge operator W_(k) under stochastic constraints,and then under the mean square error criterion,compares the new estimator with the existing weighted mixed estimator β_(WME),weighted mixed ridge estimator β_(WMRE),weighted mixed Liu estimator β_(WMLE) and the weighted mixed almost unbiased Liu estimator β_(WAULE),giving the necessary and sufficient conditions for the new estimator to be superior to other estimators.Finally,the paper verifies the theoretical results by a specific data example.The results show that the new estimator is superior to other weighted mixed estimators under appropriate parameter values.

关 键 词：多重共线性 随机约束 加权混合两参数估计 均方误差准则 

分 类 号：O212.1[理学—概率论与数理统计]