数学物理学报 ›› 2009, Vol. 29 ›› Issue (6): 1465-1476.

• 论文 •    下一篇

缺失数据下EV模型的调整最小二乘估计

  

  1. 北京师范大学数学科学学院, 北京 100875
  • 收稿日期:2007-11-20 修回日期:2008-09-13 出版日期:2009-12-25 发布日期:2009-12-25
  • 基金资助:

    国家自然科学基金(10771017)和教育部科学技术研究重大项目(309007)资助

Weighted Adjust LS Estimation |in EV Model with Missing Data

  1. School of Mathematical Science, Beijing Normal University, Beijing 100875
  • Received:2007-11-20 Revised:2008-09-13 Online:2009-12-25 Published:2009-12-25
  • Supported by:

    国家自然科学基金(10771017)和教育部科学技术研究重大项目(309007)资助

摘要:

该文考虑协变量缺失时的多元线性EV模型参数的估计, 其中协变量的缺失机制是Rubin(1976)提出的随机缺失(MAR).
利用加权调整最小二乘方法给出参数估计, 证明了估计的相合性和渐近正态性. 数值模拟结果表明所给的估计性态良好.

关键词: 线性EV模型, MAR, 加权调整最小二乘, 相合性,  渐近正态性

Abstract:

The multivariate linear errors-in-variables model when the regressors are missing at random in the sense of
Rubin (1976) is considered in this paper. The weighted adjust LS parameter estimation based on inverse probability is proposed. It is shown that the estimators have consistency and asymptotic normality.
Simulations are shown that our estimators perform well.

Key words: Linear errors-in-variables model, Weighted adjust LSestimation, Inverse probability, Consistency, Asymptotic normality

中图分类号: 

  • 62F12