Stability analysis of the Bayesian TQR model estimated by using Lasso and modified Lasso techniques

• haitham khozaee

  عربي عراقي مسلم

  Author

Abstract:

In many applied research studies that use statistical models in data analysis, the focus is often on testing the suitability of the applied model in interpreting the data and making the necessary predictions. However, this often neglects to examine the model's stability at different data levels, which can affect the reliability and consistency of these models when applied under varying conditions. The TQR model is one of the important models widely used in data analysis. This research aimed to analyze the stability of the Tobit Quantile Regression (TQR) model using Bayesian methods, Lasso techniques, and a modified Lasso technique proposed by the researcher. The focus was on examining the model's stability across quantiles. Important metrics such as Quantile Stability Across and Shrinkage Paths were used to achieve optimal results and compare the performance of these techniques. The results demonstrated the importance of these metrics combined in analyzing model stability, as they provide a dual evaluation of different quantiles and shrinkage paths. Furthermore, the results showed that Lasso techniques are more effective at maintaining model stability than the modified Lasso technique. The significance of this research lies in its presentation of improved techniques in Bayesian quantile models for analyzing high-dimensional data.

Keywords: Tobit  Quantile Regression Model, Lasso, Modified Lasso, Stability Across Quantiles ، Shrinkage Paths

• PDF (Arabic)

Copyright (c) 2026 Growth Leadership

This work is licensed under a Creative Commons Attribution 4.0 International License.

This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Authors retain the copyright and full publishing rights without restrictions.