摘要
多元线性回归模型的经典假定之一是解释变量之间不存在线性关系。但在实际应用中,多元线性回归模型中的解释变量之间往往存在近似的线性关系,如果仍然用最小二乘法估计模型,会造成分析结果不准确甚至严重偏离变量间本来的依存关系。为此,首先总结了多重共线性的检验方法,然后探讨了多重共线性常用的修正方法,最后结合实例演绎了逐步回归法和主成分回归法的具体应用,为现实经济问题中多重共线性的检验与处理提供一定借鉴。
There being no linear relationship among interpretation variables is one of the classical assumptions in multiple linear regression model.However,in the practical application,there is often an approximate linear relation.If we still use the method of ordinary least squares to estimate the model,the result may become incorrect and even far from the original relationship among the variables.Therefore,the paper first summarizes the test methods of Multicollinearity.And then,the paper summarizes the commom correction methods of multicollinearity.Finally,the application of stepwise regression and principal component regression is deduced by using an example.The research will provide some reference for the test and treatment of multicollinearity in real economic problems.
作者
刘芳
董奋义
LIU Fang;DONG Fenyi(College of In formation and Management Science,Henan Agricultural University.Zhengzhou 450046,China)
出处
《中原工学院学报》
CAS
2020年第1期44-48,55,共6页
Journal of Zhongyuan University of Technology
基金
河南省教育科学“十三五”规划课题(2018-JKGHYB-0028).
关键词
多重共线性
诊断
补救措施
逐步回归法
主成分回归
multicollinearity
diagnosis
remedial measures
stepwise regression
principal component regression
作者简介
刘芳(1974-),女,副教授,博士生,主要研究方向为计量经济学、农业经济。E-mail:gaoyt-@163.com。
