| 基于GMM的缺失数据回归模型的半参数估计 |
| |
| 引用本文： | 邓明. 基于GMM的缺失数据回归模型的半参数估计[J]. 统计与信息论坛, 2013, 28(3): 9-15 |
| |
| 作者姓名： | 邓明 |
| |
| 作者单位： | 厦门大学经济学院,福建厦门361005;中国社会科学院城市发展与环境研究所,北京100836 |
| |
| 基金项目： | 中国博士后科学基金项目《城市间土地财政的竞争外溢与房价的空间传导》，全国统计科研计划项目《时变系数的空间面板数据模型——理论与应用》 |
| |
| 摘 要： | 响应变量存在数据缺失的情况广泛出现在社会经济研究中,对响应变量存在数据缺失的回归模型提出了一个在矩估计框架下的单一的半参数估计量,这种估计量保留了参数回归估计量与非参数匹配估计量的特性,从而使得该估计量既能在响应变量被观测的子样本中保持较好的拟合性,又能够降低响应变量未被观测的子样本的估计误差,并且证明了这种估计量是一致、渐进正态估计量。
|
| 关 键 词： | 缺失数据 半参数估计 广义矩估计 |
The Semiparametric Estimation of Regression Model with Missing Data Based on GMM |
| |
| DENG Ming. The Semiparametric Estimation of Regression Model with Missing Data Based on GMM[J]. Statistics & Information Tribune, 2013, 28(3): 9-15 |
| |
| Authors: | DENG Ming |
| |
| Affiliation: | DENG Ming (1. Department of Public Economics, Xiamen University, Xiamen 361005, China;2. Institute for Urban and Environmental Studies, Chinese Academy of Social Science, Beijing 100836, China) |
| |
| Abstract: | Responding variables having missing data is a common phenomenon in the social and economic research. This paper proposes a single semiparametric estimator of the regression model in the moment method framework when the responding variables have missing data. This estimator combines the parametric regression estimator and the nonparametric matching estimator, which has good fitness in the sub sample that responding variable Y is observed and has low bias in the sub sample that responding variable Y is unobserved. This paper also proves that this estimator is consistent and asymptotic nominal. |
| |
| Keywords: | missing data semiparametric estimation GMM |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
