|
|
 |
| |
|
|
|
Modeling and Simulation of Bayesian Adaptive Lasso Variable Selection |
|
|
|
PP: 95-104 |
|
|
doi:10.18576/jsapl/120202
|
|
|
|
Author(s) |
|
|
|
A. Jassim Hassan,
A. Naeem Flaih,
H. Elsalloukh,
J. Guardiola,
|
|
|
|
Abstract |
|
|
| In this paper, we propose a new hierarchical model for the prior distribution in linear regression. The Bayesian Adaptive Lasso is investigated to explore the ability of the proposed method in both explanatory and variable selection aspects. New posterior distributions are developed based on the proposed hierarchical model. We employed the Gibbs sampler to estimate posterior means of parameters. The proposed method assumes that the variance of the data follows the Inv-x2 distribution in the hierarchical model. We conducted a simulation study to assess the performance of the proposed method under different sample sizes and compared the results with the traditional Lasso method, the Bayesian Lasso method, and the MMAD and SD criteria. The simulation was also used to assess estimation accuracy. The proposed method yielded the smallest values of MMAD and SD. |
|
|
|
|
|
|
|
