Optimized two-parameter heteroscedastic-adjusted ridge estimators for linear regression model
Article
Zaman, Q, Wasim, D, Ismail, M et al. (2025). Optimized two-parameter heteroscedastic-adjusted ridge estimators for linear regression model
. COMMUNICATIONS IN STATISTICS-THEORY AND METHODS, 10.1080/03610926.2025.2581242
Zaman, Q, Wasim, D, Ismail, M et al. (2025). Optimized two-parameter heteroscedastic-adjusted ridge estimators for linear regression model
. COMMUNICATIONS IN STATISTICS-THEORY AND METHODS, 10.1080/03610926.2025.2581242
This study addresses the development of Two-Parameter Heteroscedasticity Ridge Estimators (TPHRE) aimed at minimizing the Mean Square Error (MSE) in datasets affected by both multicollinearity and heteroscedasticity. Over the past two decades, two-parameter ridge regression has become a widely adopted approach for mitigating multicollinearity, with the ridge parameter k, playing a crucial role in optimization. In this paper, we propose a novel scaling factor that substantially improves the performance of ridge estimators compared to traditional approaches. Our proposed estimator, TPHRE (IQW1), demonstrates superior capability in handling heteroscedasticity, providing more reliable and efficient estimates. Simulation studies confirm its improved performance under high multicollinearity and heteroscedasticity, while a real-world application further highlights its practical advantages in analyzing data with collinear predictors and heteroscedastic errors.
