摘要
文章针对面板数据复合分位数回归模型,提出了模型中回归系数的估计方法。首先通过乘一个幂等矩阵消除个体固定效应的影响,避免了估计个体固定效应项;然后利用复合分位数回归方法估计回归系数。在一些正则条件下,证明了复合分位数回归估计的渐近性质。
This paper proposes an estimation method of the regression coefficient for composite quantile regression model with panel data. Firstly, the paper eliminates the impact of individual fixed effect by multiplying an idempotent matrix so as to avoid individual fixed effect. And then the paper uses the composite quantile regression method to estimate the regression coefficient. Under some regularity conditions, the composite quantile regression estimation is proved to be asymptotically normal.
作者
徐洁
杨宜平
Xu Jie,Yang Yiping(School of Mathematics and Statistics, Chongqing Technology and Business University, Chongqing 400067, Chin)
出处
《统计与决策》
CSSCI
北大核心
2018年第5期19-21,共3页
Statistics & Decision
基金
国家自然科学基金资助项目(11301569)
重庆市基础与前沿研究计划项目(cstc2015jcyj A00023)
重庆市教委科学技术研究项目(k J1500614)
关键词
复合分位数回归
面板数据
最小二乘法
渐近正态
composite quantile regression
panel data
least squares method
asymptotically normal
