Special ridge-type estimator: Simulation and application to chemical data

This study delves into regularization techniques, such as ridge regression, Liu estimator, and Kibria–Lukman estimator, as alternatives to the maximum likelihood method for addressing multicollinearity in beta regression models. While the existing literature presents various shrinkage parameters, this research identifies a crucial gap in determining the optimal ridge parameter. Introducing a groundbreaking approach—ridge regression without a shrinkage parameter—we showcase its theoretical promise through analyses and simulations. The proposed estimator demonstrates competitive performance against established methods, outperforming Liu and maximum likelihood estimators.