Fixed and Random Effect Selection in Finite Mixture of Linear Mixed-Effects Models

• Activity Number: 636
• Type: Invited
• Date/Time: Thursday, August 2, 2012 : 10:30 AM to 12:20 PM
• Sponsor: SSC
• Author(s): Abbas Khalili*+ and Yeting Du and Johanna Neslehova and Russel Steele
• Companies: McGill University and McGill University and McGill University and McGill University
• Keywords: Linear mixed effects models ; Mixture models ; Penalized likelihood methods ; LASSO ; SCAD

Abstract

Linear mixed effects (LME) models are frequently used for modeling longitudinal data. One complicating factor in analyzing such data is that samples are sometimes obtained from a population with significant underlying heterogeneity, which is hard to capture by a single LME model. Such problems may be addressed by a finite mixture of linear mixed-effects (FMLME) models, which segments the population into subpopulations and models each subpopulation by a distinct LME model. In the analysis of such models, a primary objective is to assess significant fixed effects and/or random effects that could vary across different components of the mixture model. Classical all-subset selection techniques are computationally expensive as the number of covariates and components in the mixture model increases. In this article, we introduce a new penalized likelihood approach for mixed effects selection in the FMLME models. Performance of the new method is studied asymptotically and also through extensive simulations. We illustrate the usage of the new method by analyzing a dataset on patients enrolled in the Canadian Scleroderma Research Group Registry.