含有测量误差的Panel数据模型的统计推断  

Statistical inference for the Panel data model with errors-in-variables

作　　者：岳莉莉 史建红[2] 李高荣 YUE LiLi SHI JianHong LI GaoRong

机构地区：[1]北京工业大学应用数理学院,北京100124 [2]山西师范大学数学与计算机科学学院,临汾041004 [3]北京工业大学北京科学与工程计算研究院,北京100124

出　　处：《中国科学：数学》2017年第9期1077-1088,共12页Scientia Sinica：Mathematica

基　　金：国家自然科学基金(批准号:11471029);北京市自然科学基金(批准号:1142002);山西省自然科学基金(批准号:2013011002-1)资助项目

摘　　要：本文考虑含有测量误差的Panel数据模型在约束条件下的参数估计和检验问题.首先,给出未知参数向量的纠正最小二乘估计,在一些正则条件下研究其渐近性质.其次,对线性约束条件的假设检验问题,利用原假设和备择假设下纠正残差平方和的差构造检验统计量,并证明该检验统计量的渐近分布.最后,通过数值模拟研究验证所提估计和检验方法的有效性.In this paper, the Panel data model with errors-in-variables is considered, and the parameter es- timation and testing problem of restricted condition are studied. Firstly, a corrected least squares method is proposed to estimate the unknown parameter vector, and the asymptotic property of the obtained estimator is shown under some regularity conditions. Then, to test the validity of linear restriction, a test statistic based on the difference of the corrected residual sum of squares under the null and alternative hypotheses is constructed, and its limiting distribution is derived. The effectiveness of the proposed methods is assessed by some simulation studies.

关 键 词：Panel数据模型 测量误差 纠正最小二乘 约束估计 渐近性质 

分 类 号：O212.1[理学—概率论与数理统计]