A novel heteroscedasticity-adjusted two-parameter ridge estimator in linear regression
Abstract
Ordinary least squares (OLS) regression relies on key assumptions, including uncorrelated regressors and homoscedastic errors. However, data from various fields of study have become more complex and typically do not meet these assumptions. Penalized regression techniques are vital for achieving reliable results when these assumptions are not satisfied. Therefore, this paper proposes a novel two-parameter ridge estimator as a solution to address both multicollinearity and heteroscedasticity. Utilizing the mean square error criterion, our comprehensive simulations and real-world applications showcase the superior performance of the proposed method, especially in complex data scenarios where traditional approaches fail.
Fields of Study & Subjects
Advanced Statistical Methods and Models
Statistical Methods and Inference
Statistical and numerical algorithms
Publication details
- Journal / source
- Journal of Statistical Computation and Simulation
- Volume
- 96
- Issue
- 9
- Pages
- 2106–2125
- Publication date
- 13 January 2026
- Language
- EN