1: \begin{abstract}

2: Linear Mixed-Effects (LME) models are a fundamental tool for modeling clustered data, including cohort studies, longitudinal data analysis, and meta-analysis.  

3: The design and analysis of variable selection methods for LMEs is 

4: considerably more difficult than for linear regression because LME models are nonlinear. 

5: The approach considered here is motivated by a recent method for sparse relaxed regularized regression (SR3) for variable selection in the context of linear regression.

6: The theoretical underpinnings 

7: for the proposed extension to LMEs are developed, including consistency results, variational properties, 

8: implementability of optimization methods, and convergence results.  

9: In particular we provide convergence analyses for a basic implementation of SR3 for LME (called MSR3) and an accelerated 

10: hybred algorithm (called MSR3-fast).  

11: Numerical results show the utility and speed of these algorithms on realistic simulated datasets. 

12: The numerical implementations are available  

13: in an open source \texttt{python} package \texttt{pysr3}.

14: \end{abstract}

15: