leverage and influential observations on the restricted liu estimator in the linear mixed measurement error models

A fundamental step in constructing linear regression models is the identification of atypical observations, specifically high-leverage and influential points. the coexistence of atypical observations and severe multicollinearity complicates statistical analysis and reduces the effectiveness of conventional diagnostic methods. to address these challenges, this paper proposes diagnostic methods based on the restricted liu estimator for identifying atypical observations in linear mixed measurement error models. specifically, generalized leverage matrices are derived using the restricted liu estimator to identify high-leverage points. furthermore, several case-deletion measures are developed based on the restricted liu estimator to serve as robust tools for diagnosing influential points. the performance of the proposed methods is evaluated through two simulation studies and an application to a real-life example.