Efficient Leave-One-Out Cross-Validation and WAIC for Bayesian Models

Efficient approximate leave-one-out cross-validation (LOO) for Bayesian models fit using Markov chain Monte Carlo, as described in Vehtari, Gelman, and Gabry (2017) . The approximation uses Pareto smoothed importance sampling (PSIS), a new procedure for regularizing importance weights. As a byproduct of the calculations, we also obtain approximate standard errors for estimated predictive errors and for the comparison of predictive errors between models. The package also provides methods for using stacking and other model weighting techniques to average Bayesian predictive distributions.

Bayesian Inference with Laplace Approximations and P-Splines

Laplace approximations and penalized B-splines are combined for fast Bayesian inference in latent Gaussian models. The routines can be used to fit survival models, especially proportional hazards and promotion time cure models (Gressani, O. and Lambert, P. (2018) ). The Laplace-P-spline methodology can also be implemented for inference in (generalized) additive models (Gressani, O. and Lambert, P. (2021) ). See the associated website for more information and examples.

Bayesian Monotonic Regression Using a Marked Point Process Construction

An extended version of the nonparametric Bayesian monotonic regression procedure described in Saarela & Arjas (2011) , allowing for multiple additive monotonic components in the linear predictor, and time-to-event outcomes through case-base sampling. The extension and its applications, including estimation of absolute risks, are described in Saarela & Arjas (2015) . The package also implements the nonparametric ordinal regression model described in Saarela, Rohrbeck & Arjas .

Longitudinal Gaussian Process Regression

Interpretable nonparametric modeling of longitudinal data using additive Gaussian process regression. Contains functionality for inferring covariate effects and assessing covariate relevances. Models are specified using a convenient formula syntax, and can include shared, group-specific, non-stationary, heterogeneous and temporally uncertain effects. Bayesian inference for model parameters is performed using 'Stan'. The modeling approach and methods are described in detail in Timonen et al. (2021) .

Complete Environment for Bayesian Inference

Provides a complete environment for Bayesian inference using a variety of different samplers (see ?LaplacesDemon for an overview).

Robust Bayesian Variable Selection via Expectation-Maximization

Variable selection methods have been extensively developed for analyzing highdimensional omics data within both the frequentist and Bayesian frameworks. This package provides implementations of the spike-and-slab quantile (group) LASSO which have been developed along the line of Bayesian hierarchical models but deeply rooted in frequentist regularization methods by utilizing Expectation–Maximization (EM) algorithm. The spike-and-slab quantile LASSO can handle data irregularity in terms of skewness and outliers in response variables, compared to its non-robust alternative, the spike-and-slab LASSO, which has also been implemented in the package. In addition, procedures for fitting the spike-and-slab quantile group LASSO and its non-robust counterpart have been implemented in the form of quantile/least-square varying coefficient mixed effect models for high-dimensional longitudinal data. The core module of this package is developed in 'C++'.

Empirical Bayesian Tobit Matrix Estimation

Estimation tools for multidimensional Gaussian means using empirical Bayesian g-modeling. Methods are able to handle fully observed data as well as left-, right-, and interval-censored observations (Tobit likelihood); descriptions of these methods can be found in Barbehenn and Zhao (2023) . Additional, lower-level functionality based on Kiefer and Wolfowitz (1956) and Jiang and Zhang (2009) is provided that can be used to accelerate many empirical Bayes and nonparametric maximum likelihood problems.

Dynamic Stochastic General Equilibrium Models

Specify, solve, and estimate dynamic stochastic general equilibrium (DSGE) models by maximum likelihood and Bayesian methods. Supports both linear models via an equation-based formula interface and nonlinear models via string-based equations with first-order perturbation (linearization around deterministic steady state). Solution uses the method of undetermined coefficients (Klein, 2000 ). Likelihood evaluated via the Kalman filter. Bayesian estimation uses adaptive Random-Walk Metropolis-Hastings with prior specification. Additional tools include Kalman smoothing, historical shock decomposition, local identification diagnostics, parameter sensitivity analysis, second-order perturbation, occasionally binding constraints, impulse-response functions, forecasting, and robust standard errors.

A 'shiny' Application for Network Meta-Analysis

Conduct network meta-analyses through a graphical user interface using 'bnma', 'gemtc' and 'netmeta' with additional analysis provided by 'meta' and 'metafor'. Frequentist, Bayesian, meta-regression and baseline risk meta-regression analyses can all be conducted using a consistent data structure and terminology. Many options are provided for downloading publication-ready outputs and analyses can be reproduced outside of the application by downloading a 'quarto' file. The interface was generated using 'shinyscholar'. The initial version of the app was described by Owen et al. (2018) , Bayesian ranking visualisations were described by Nevill et al. (2023) and metaregression was described by Morris et al. (2025) .

An Implementation of the Bayesian Markov (Renewal) Mixed Models

The Bayesian Markov renewal mixed models take sequentially observed categorical data with continuous duration times, being either state duration or inter-state duration. These models comprehensively analyze the stochastic dynamics of both state transitions and duration times under the influence of multiple exogenous factors and random individual effect. The default setting flexibly models the transition probabilities using Dirichlet mixtures and the duration times using gamma mixtures. It also provides the flexibility of modeling the categorical sequences using Bayesian Markov mixed models alone, either ignoring the duration times altogether or dividing duration time into multiples of an additional category in the sequence by a user-specific unit. The package allows extensive inference of the state transition probabilities and the duration times as well as relevant plots and graphs. It also includes a synthetic data set to demonstrate the desired format of input data set and the utility of various functions. Methods for Bayesian Markov renewal mixed models are as described in: Abhra Sarkar et al., (2018) and Yutong Wu et al., (2022) .