明瑟收入函数模型变量选择方法比较研究——来自云南农村住户数据的实例

A Comparative Study of Variable Selection under Mincer Income Function Model: Evidence from Data of Rural Households of Yunnan Province

本文基于2010年云南红河哈尼族彝族自治州农村3000户数据库,比较研究了压缩系数法(岭回归、LASSO、SCAD和适应性LASSO)在农户明瑟收入函数模型中的变量选择问题。结果发现:SCAD方法优于其他压缩系数法,对相关应用研究具有借鉴意义。

This study compared the Shrinkage methods(Ridge regression, Lasso, SCAD and Adaptive Lasso) and applied them into variables selection under Mincer income function model, which is based on the data of 3000 rural households of Yunnan Honghe Hani and Yi Autonomous Prefecture in 2010. The results showed, the SCAD method is much more effective than the other Shrinkage methods; the results have reference to the relevant application studies.