Fully Efficient and Outlier-Robust Estimation in the Linear Mixed Model

• Activity Number: 73
• Type: Contributed
• Date/Time: Sunday, July 31, 2016 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistical Learning and Data Science
• Author(s): Won Gyo Suh* and Howard Bondell
• Companies: North Carolina State University and North Carolina State University
• Keywords: linear mixed-effects model ; robust ; outlier ; efficient ; mixed model ; empirical likelihood

Abstract

Since Linear mixed-effects models provide general and flexible approach to correlated data, they are widely used. Outliers in Linear mixed-effects model can be more problematic than in Linear model because Linear mixed-effects model has more source of variance. Large- and finite-sample efficiency and resistance to outliers are the key goals of robust statistics. Usually both are hard to achieve at the same time however Howard D. Bondell and Leonard A. Stefanski suggested robust estimation in linear model not only efficient but also robust in their paper (2013). Specifically, by using empirical likelihood efficiency of estimations is obtained and by constraining the associated sum of weighted studentized residual squares, robustness properties are obtained. In this paper, I will apply the idea to Linear mixed-effects models and compare efficiency and robustness with current robust estimator.