A Comparison of Some Penalized Methods for Simultaneous Estimation and Variable Selection of the Linear Regression Model under Multicollinearity:A Simulation Study

Document Type : Original Article

Author

Statistics,Mathematics and insurance department Faculty of Commerce Alexandria University Alexandria Egypt

Abstract

The research introduces a comparison of some penalized methods for simultaneous variable selection and estimation of the linear regression model under  multicollinearity. A simulation study was conducted to compare the performance of these methods including 120 different situations resulting from interaction of four factors: sample size, random error variance, degree of linear correlation among explanatory variables, and regression parameters of the true model. The simulation study shows that SEALASSO method outperforms SCAD and MCP methods in terms of percentage of selecting the true model, and competes favorably with them in terms of estimation accuracy

Keywords

Files

Share

How to cite

Statistics