摘要
与普通最小二乘法相比,分位数回归能够更充分反映自变量对不同部分因变量的分布产生不同的影响,有着十分广泛的应用。本文对分位数回归的思想做了一个简单的介绍,并将其方法应用于恩格尔定律中,比较分析了异方差和同方差下分位数回归与普通最小二乘法的优劣。
Compared with ordinary least squares, quantile regression can more fully reflect that the dependent variable has different effects in the different parts of the distribution of the independent variables, and has a very wide range of applications. The paper makes a brief introduction to the idea of quantile regression, applying the methods into Engel's law and having a comparative analysis of the good or bad about quantile regression and ordinary least squares under different variance and same variance.
出处
《统计教育》
2009年第10期58-61,共4页
Statistical education