均值方差联合模型的SEE变量选择

SEE Variable Selection for Joint Mean and Variance Models

摘要: 对方差建立回归模型分析是处理异方差问题中最常用的方法之一。本文基于均值方差联合模型，结合光滑阈估计方程(Smooth Threshold Estimating Equation，简记SEE)方法研究该模型的变量选择方法。该变量选择方法可以同时进行参数估计和变量选择，并且不需要解任何凸优化问题，因此实际应用中将大大减少计算量。最后, 通过随机模拟实验验证了所提出方法的有效性与可行性。

Abstract: The method based on modeling the variance is one of the most commonly used methods to deal with heteroscedasticity. In this paper, we propose a variable selection procedure based on the smooth threshold estimating equations for joint mean and variance models. The proposed variable selection method can select variables and estimate coefficients simultaneously, and does not need to solve convex optimization problem so as to largely reduce computation quantity in practice. Finally, we make some simulations to show that the proposed procedure works satisfactorily.