关于空间回归模型残差项分布未知的研究

Research on Spatial Regression Model with Unknown Distribution Residuals

文章分析了剖面似然方法,在残差项分布未知的条件下研究了空间回归模型的回归系数的估计问题,并在模拟实验中与二阶段最小二乘方法做比较,得出剖面似然方法所得到的参数估计更加稳健的结论。

This paper analyzed the profile likelihood method, mainly studied the estimation of regression coefficients under the condition of unknown distribution of residuals of spatial regression model and made some simulation experiments comparing with two-stage least squares method. The simulation obtained the conclusions that the parameter estimated by profile likelihood method demonstrated more robust.