统计研究 ›› 2010, Vol. 27 ›› Issue (12): 16-22.

• 论文 • 上一篇    下一篇

基于稳健MM估计的统计数据质量评估方法

卢二坡 黄炳艺   

  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2010-12-15 发布日期:2010-12-15

An Evaluation Method of Statistical Data Quality Based on Robust MM Estimator

Lu Erpo & Huang Bingyi   

  • Received:1900-01-01 Revised:1900-01-01 Online:2010-12-15 Published:2010-12-15

摘要: 政府统计数据质量是当前各界关注的热点问题,如何采用严谨的诊断方法,对我国统计数据进行科学的评估具有重要的现实意义。稳健回归方法可使求出的回归估计不受异常值的强烈影响,并且能更好的识别异常点。本文首次运用基于稳健MM估计的异常值诊断方法,在生产函数模型的框架下,分别使用两种不同的劳动投入数据,对改革以来我国GDP数据质量进行了评估。结果表明,基于稳健MM估计的异常值诊断方法可有效的解决传统方法容易出现的多个异常点的掩盖现象,改革以来我国的GDP数据是相对可靠的。

关键词: 统计数据质量, 稳健MM估计, 异常值诊断

Abstract: Statistical data quality is a widespread concern issue. It is significant to evaluate China’s statistics scientifically using rigorous diagnostic methods. Robust regression is resistant to the influence of outliers and can be used as a good tool identifying outliers. Under the framework of production function model and using two different labor input data, this paper first applies robust MM estimator as a detection tool of outliers to assess China’s GDP data quality since reform. The results show that the problem of masking effect which exists in traditional methods can be effectively solved and China’s GDP data since reform is relatively reliable.

Key words: Statistical data quality, Robust MM estimator, Outliers detection