Estimate (Generalized) Linear Mixed Models with Factor Structures

Utilizes the 'lme4' and 'optimx' packages (previously the optim() function from 'stats') to estimate (generalized) linear mixed models (GLMM) with factor structures using a profile likelihood approach, as outlined in Jeon and Rabe-Hesketh (2012) and Rockwood and Jeon (2019) . Factor analysis and item response models can be extended to allow for an arbitrary number of nested and crossed random effects, making it useful for multilevel and cross-classified models.

Penalized ECME Estimation for Censored Linear Mixed Models

Fits Gaussian linear mixed models with a random intercept when the response is subject to left, right, and/or interval censoring, using the Expectation/Conditional Maximization Either (ECME) algorithm of Liu and Rubin (1994) in the spirit of the fast censored-response mixed-model algorithm of Vaida and Liu (2009). Simultaneous estimation and variable selection is supported through coordinate-descent penalized maximization with Lasso, Adaptive Lasso, SCAD, MCP, Elastic Net, and Ridge penalties (no penalty is also supported). The random intercept is integrated out by Gauss-Hermite quadrature at every iteration, and the two ECME conditional-maximization steps respectively maximize the expected penalized complete-data objective (for the regression coefficients) and the actual observed-data marginal likelihood (for the variance components), which is the defining feature of ECME relative to plain ECM/EM. The package provides a single-fit engine, a sequential/parallel penalty-parameter grid search with information-criterion or cross-validated selection, data-dependent or user-supplied lambda grids, and an Expectation-Maximization based treatment of a completely missing (at random) response, sharing the same truncated-normal machinery used for censoring. References: Liu and Rubin (1994) "The ECME algorithm: A simple extension of EM and ECM with faster monotone convergence" ; Vaida and Liu (2009) "Fast Implementation for Normal Mixed Effects Models With Censored Response" .

Robust Bayesian Longitudinal Regularized Semiparametric Mixed Models

Our recently developed fully robust Bayesian semiparametric mixed-effect model for high-dimensional longitudinal studies with heterogeneous observations can be implemented through this package. This model can distinguish between time-varying interactions and constant-effect-only cases to avoid model misspecifications. Facilitated by spike-and-slab priors, this model leads to superior performance in estimation, identification and statistical inference. In particular, robust Bayesian inferences in terms of valid Bayesian credible intervals on both parametric and nonparametric effects can be validated on finite samples. The Markov chain Monte Carlo algorithms of the proposed and alternative models are efficiently implemented in 'C++'.

Generalized Fiducial Inference for Normal Linear Mixed Models

Simulation of the generalized fiducial distribution for normal linear mixed models with interval data. Fiducial inference is somehow similar to Bayesian inference, in the sense that it is based on a distribution that represents the uncertainty about the parameters, like the posterior distribution in Bayesian statistics. It does not require a prior distribution, and it yields results close to frequentist results. Reference: Cisewski and Hannig (2012) .

Variable Selection in Linear Mixed Models for SNP Data

Fit penalized multivariable linear mixed models with a single random effect to control for population structure in genetic association studies. The goal is to simultaneously fit many genetic variants at the same time, in order to select markers that are independently associated with the response. Can also handle prior annotation information, for example, rare variants, in the form of variable weights. For more information, see the website below and the accompanying paper: Bhatnagar et al., "Simultaneous SNP selection and adjustment for population structure in high dimensional prediction models", 2020, .

Tables and Graphs for Mixed Models for Repeated Measures (MMRM)

Mixed models for repeated measures (MMRM) are a popular choice for analyzing longitudinal continuous outcomes in randomized clinical trials and beyond; see for example Cnaan, Laird and Slasor (1997) . This package provides an interface for fitting MMRM within the 'tern' < https://cran.r-project.org/package=tern> framework by Zhu et al. (2023) and tabulate results easily using 'rtables' < https://cran.r-project.org/package=rtables> by Becker et al. (2023). It builds on 'mmrm' < https://cran.r-project.org/package=mmrm> by Sabanés Bové et al. (2023) for the actual MMRM computations.

Bayesian Spatial and Space-Time Linear Mixed Models

Fits Bayesian linear mixed models for spatial and space-time data with fixed effects, independent and identically distributed (iid) grouped random effects, and structured latent processes. The formula interface supports first-order autoregressive (AR(1)) effects, dense Gaussian processes, nearest-neighbor Gaussian processes, proper and Leroux conditional autoregressive (CAR) effects, ordered directed acyclic graph autoregressive (DAGAR) effects, separable CAR-time and DAGAR-time effects, and spatially varying coefficients. The sampler uses sparse precision matrix calculations when available and includes post-fitting tools for latent process recovery, fitted values, prediction, pointwise log likelihoods, and posterior sample extraction. Method details include Datta et al. (2016) , Finley et al. (2019) , Datta et al. (2019) , and May and Finley (2025) .

Bayesian Robust Generalized Mixed Models for Longitudinal Data

To perform model estimation using MCMC algorithms with Bayesian methods for incomplete longitudinal studies on binary and ordinal outcomes that are measured repeatedly on subjects over time with drop-outs. Details about the method can be found in the vignette or < https://sites.google.com/view/kuojunglee/r-packages/bayesrgmm>.

Power Analysis for Linear and Generalized Linear Mixed Models

Prospective power analysis and design planning using the "plug-in" approach described by Piepho et al. (2022) and Stroup, Ptukhina and Garai (2024) . Expected treatment means and the proposed experimental design are encoded in an artificial data set, and a model is "fit" to it with random effect variances, correlations and residual dispersion fixed at user-specified values rather than being estimated. The resulting standard errors are used to compute the power of F-tests and user-defined contrasts, as well as Type S (sign) and Type M (magnitude) error rates (Gelman and Carlin, 2014 ). Models are fit with 'glmmTMB', so the approach accommodates non-normal responses, heteroscedasticity, nested and crossed random effects, correlated random slopes, and structured covariances such as AR(1) and spatial correlation. Helper functions convert standard deviations and correlations into the internal parameterization of 'glmmTMB' and report the order in which random effect parameters must be supplied. Denominator degrees of freedom can be computed by the Kenward-Roger, residual or asymptotic methods (or can be specified by users), and contrasts are specified through 'emmeans'. Vignettes cover estimated marginal means, contrast specification, and worked examples for common agricultural designs.

Fit a Cosinor Model Using a Generalized Mixed Modeling Framework

Allows users to fit a cosinor model using the 'glmmTMB' framework. This extends on existing cosinor modeling packages, including 'cosinor' and 'circacompare', by including a wide range of available link functions and the capability to fit mixed models. The cosinor model is described by Cornelissen (2014) .