A New Quantile-Based Approach for LASSO Estimation
Regularization regression techniques are widely used to overcome a model’s parameter estimation problem in the presence of multicollinearity. Several biased techniques are available in the literature, including ridge, Least Angle Shrinkage Selection Operator (LASSO), and elastic net. In this work, w...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-03-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/11/6/1452 |
| _version_ | 1797610285861699584 |
|---|---|
| author | Ismail Shah Hina Naz Sajid Ali Amani Almohaimeed Showkat Ahmad Lone |
| author_facet | Ismail Shah Hina Naz Sajid Ali Amani Almohaimeed Showkat Ahmad Lone |
| author_sort | Ismail Shah |
| collection | DOAJ |
| description | Regularization regression techniques are widely used to overcome a model’s parameter estimation problem in the presence of multicollinearity. Several biased techniques are available in the literature, including ridge, Least Angle Shrinkage Selection Operator (LASSO), and elastic net. In this work, we study the performance of the classical LASSO, adaptive LASSO, and ordinary least squares (OLS) methods in high-multicollinearity scenarios and propose some new estimators for estimating the LASSO parameter “k”. The performance of the proposed estimators is evaluated using extensive Monte Carlo simulations and real-life examples. Based on the mean square error criterion, the results suggest that the proposed estimators outperformed the existing estimators. |
| first_indexed | 2024-03-11T06:13:17Z |
| format | Article |
| id | doaj.art-5691d89cac214d1888b8056135da9a2d |
| institution | Directory Open Access Journal |
| issn | 2227-7390 |
| language | English |
| last_indexed | 2024-03-11T06:13:17Z |
| publishDate | 2023-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Mathematics |
| spelling | doaj.art-5691d89cac214d1888b8056135da9a2d2023-11-17T12:28:50ZengMDPI AGMathematics2227-73902023-03-01116145210.3390/math11061452A New Quantile-Based Approach for LASSO EstimationIsmail Shah0Hina Naz1Sajid Ali2Amani Almohaimeed3Showkat Ahmad Lone4Department of Statistics, Quaid-i-Azam University, Islamabad 45320, PakistanDepartment of Statistics, Quaid-i-Azam University, Islamabad 45320, PakistanDepartment of Statistics, Quaid-i-Azam University, Islamabad 45320, PakistanDepartment of Statistics and Operation Research, College of Science, Qassim University, Buraydah 51482, Saudi ArabiaDepartment of Basic Sciences, College of Science and Theoretical Studies, Saudi Electronic University, Riyadh 11673, Saudi ArabiaRegularization regression techniques are widely used to overcome a model’s parameter estimation problem in the presence of multicollinearity. Several biased techniques are available in the literature, including ridge, Least Angle Shrinkage Selection Operator (LASSO), and elastic net. In this work, we study the performance of the classical LASSO, adaptive LASSO, and ordinary least squares (OLS) methods in high-multicollinearity scenarios and propose some new estimators for estimating the LASSO parameter “k”. The performance of the proposed estimators is evaluated using extensive Monte Carlo simulations and real-life examples. Based on the mean square error criterion, the results suggest that the proposed estimators outperformed the existing estimators.https://www.mdpi.com/2227-7390/11/6/1452LASSOregularization methodsmulticollinearityhigh-dimensional dataMonte Carlo |
| spellingShingle | Ismail Shah Hina Naz Sajid Ali Amani Almohaimeed Showkat Ahmad Lone A New Quantile-Based Approach for LASSO Estimation Mathematics LASSO regularization methods multicollinearity high-dimensional data Monte Carlo |
| title | A New Quantile-Based Approach for LASSO Estimation |
| title_full | A New Quantile-Based Approach for LASSO Estimation |
| title_fullStr | A New Quantile-Based Approach for LASSO Estimation |
| title_full_unstemmed | A New Quantile-Based Approach for LASSO Estimation |
| title_short | A New Quantile-Based Approach for LASSO Estimation |
| title_sort | new quantile based approach for lasso estimation |
| topic | LASSO regularization methods multicollinearity high-dimensional data Monte Carlo |
| url | https://www.mdpi.com/2227-7390/11/6/1452 |
| work_keys_str_mv | AT ismailshah anewquantilebasedapproachforlassoestimation AT hinanaz anewquantilebasedapproachforlassoestimation AT sajidali anewquantilebasedapproachforlassoestimation AT amanialmohaimeed anewquantilebasedapproachforlassoestimation AT showkatahmadlone anewquantilebasedapproachforlassoestimation AT ismailshah newquantilebasedapproachforlassoestimation AT hinanaz newquantilebasedapproachforlassoestimation AT sajidali newquantilebasedapproachforlassoestimation AT amanialmohaimeed newquantilebasedapproachforlassoestimation AT showkatahmadlone newquantilebasedapproachforlassoestimation |