摘要
结合广义半参数混合效应面板模型与广义半参数时间序列模型,文章构建广义半参数趋势混合效应面板模型,考虑到解释变量的非平稳性会影响到模型的参数估计,所以将解释变量的非线性时间趋势估计出来,解释变量用非线性时间趋势代替,最后采用加权极大似然估计方法估计模型参数,通过假定被解释变量为指数分布的情况下,验证了估计方法的有效性。
Combining generalized semi-parametric mixed effect panel model with generalized semi-parametric time series model, this paper constructs a generalized semi-parametric trending mixed effect panel model. Considering that the non-stationary characteristic of explanatory variable will affect the parameter estimation, the paper estimates the nonlinear time trend of explanatory variables, and uses nonlinear time trend to replace explanatory variables. And finally the paper adopts weighted maximum likelihood estimation method to estimate model parameters and verifies the effectiveness of the proposed estimation method by supposing explanatory variables to be under the exponential distribution.
出处
《统计与决策》
CSSCI
北大核心
2018年第4期5-10,共6页
Statistics & Decision
关键词
广义半参数面板模型
加权极大似然估计
指数分布
generalized semi-parametric panel model
weighted maximum likelihood estimation
exponential distribution
