On using the penalized regression estimators to solve the multicollinearity problem
1 Department of Statistics, Faculty of Science, University of Ajdabia, Ajdabia, Libya.
2 Department of Statistics, Faculty of Science, University of Benghazi, Benghazi, Libya.
Review Article
World Journal of Advanced Research and Reviews, 2024, 24(01), 2654–2664
Publication history:
Received on 16 September 2024; revised on 26 October 2024; accepted on 29 October 2024
Abstract:
The paper compares coefficient parameter estimation efficiency using penalized regression approaches. Five estimators are employed: Ridge Regression, LASSO regression, Elastic Net (ENET) Regression, Adaptive Lasso (ALASSO) regression, and Adaptive Elastic Net (AENET) regression methods. The study uses a multiple linear regression model to address multicollinearity issues. The comparison is based on average mean square errors (MSE) using simulated data with varying sizes, numbers of independent variables, and correlation coefficients. The results are expected to be useful and will be applied to real data to determine the best-performing estimator.
Keywords:
Ridge Regression; LASSO Regression; Elastic Net Regression; Adaptive Lasso Regression; Adaptive Elastic Net Regression methods
Full text article in PDF:
Copyright information:
Copyright © 2024 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0