Extensible Package for Parallel, Batch Training of Base Learners for Ensemble Modeling

Extensible S4 classes and methods for batch training of regression and classification algorithms such as Random Forest, Gradient Boosting Machine, Neural Network, Support Vector Machines, K-Nearest Neighbors, Penalized Regression (L1/L2), and Bayesian Additive Regression Trees. These algorithms constitute a set of 'base learners', which can subsequently be combined together to form ensemble predictions. This package provides cross-validation wrappers to allow for downstream application of ensemble integration techniques, including best-error selection. All base learner estimation objects are retained, allowing for repeated prediction calls without the need for re-training. For large problems, an option is provided to save estimation objects to disk, along with prediction methods that utilize these objects. This allows users to train and predict with large ensembles of base learners without being constrained by system RAM.

Estimate Causal Effects with Borrowing Between Data Sources

Estimate population average treatment effects from a primary data source with borrowing from supplemental sources. Causal estimation is done with either a Bayesian linear model or with Bayesian additive regression trees (BART) to adjust for confounding. Borrowing is done with multisource exchangeability models (MEMs). For information on BART, see Chipman, George, & McCulloch (2010) . For information on MEMs, see Kaizer, Koopmeiners, & Hobbs (2018) .

Causal Inference for Multiple Treatments with a Binary Outcome

Different methods to conduct causal inference for multiple treatments with a binary outcome, including regression adjustment, vector matching, Bayesian additive regression trees, targeted maximum likelihood and inverse probability of treatment weighting using different generalized propensity score models such as multinomial logistic regression, generalized boosted models and super learner. For more details, see the paper by Hu et al. .

Propensity Score Predictive Inference for Generalizability

Provides a suite of Propensity Score Predictive Inference (PSPI) methods to generalize treatment effects in trials to target populations. The package includes an existing model Bayesian Causal Forest (BCF) and four PSPI models (BCF-PS, FullBART, SplineBART, DSplineBART). These methods leverage Bayesian Additive Regression Trees (BART) to adjust for high-dimensional covariates and nonlinear associations, while SplineBART and DSplineBART further use propensity score based splines to address covariate shift between trial data and target population.

Sparse Varying Coefficient BART with Global-Local Priors"

Fits sparse linear varying coefficient models (VCMs), which assert a linear relationship between an outcome and several covariates that is allowed to change as functions of additional variables known as effect modifiers. Designed for high-dimensional settings where the number of covariates (i.e., number of slopes) is comparable to or larger than the number of observations. Approximates the coefficient functions using a version of Bayesian Additive Regression Trees that can perform global-local shrinkage. For more details see Ghosh, Bhogale, and Deshpande (2026+) .

Smooth Additive Quantile Regression Models

Smooth additive quantile regression models, fitted using the methods of Fasiolo et al. (2020) . See Fasiolo at al. (2021) for an introduction to the package. Differently from 'quantreg', the smoothing parameters are estimated automatically by marginal loss minimization, while the regression coefficients are estimated using either PIRLS or Newton algorithm. The learning rate is determined so that the Bayesian credible intervals of the estimated effects have approximately the correct coverage. The main function is qgam() which is similar to gam() in 'mgcv', but fits non-parametric quantile regression models.

A Flexible Approach for Causal Inference with Multiple Treatments and Clustered Survival Outcomes

Random-intercept accelerated failure time (AFT) model utilizing Bayesian additive regression trees (BART) for drawing causal inferences about multiple treatments while accounting for the multilevel survival data structure. It also includes an interpretable sensitivity analysis approach to evaluate how the drawn causal conclusions might be altered in response to the potential magnitude of departure from the no unmeasured confounding assumption.This package implements the methods described by Hu et al. (2022) .

Bayesian Trees for Conditional Mean and Variance

A model of the form Y = f(x) + s(x) Z is fit where functions f and s are modeled with ensembles of trees and Z is standard normal. This model is developed in the paper 'Heteroscedastic BART Via Multiplicative Regression Trees' (Pratola, Chipman, George, and McCulloch, 2019, ). BART refers to Bayesian Additive Regression Trees. See the R-package 'BART'. The predictor vector x may be high dimensional. A Markov Chain Monte Carlo (MCMC) algorithm provides Bayesian posterior uncertainty for both f and s. The MCMC uses the recent innovations in Efficient Metropolis--Hastings proposal mechanisms for Bayesian regression tree models (Pratola, 2015, Bayesian Analysis, ).

Bayesian Optimization and Model-Based Optimization of Expensive Black-Box Functions

Flexible and comprehensive R toolbox for model-based optimization ('MBO'), also known as Bayesian optimization. It implements the Efficient Global Optimization Algorithm and is designed for both single- and multi- objective optimization with mixed continuous, categorical and conditional parameters. The machine learning toolbox 'mlr' provide dozens of regression learners to model the performance of the target algorithm with respect to the parameter settings. It provides many different infill criteria to guide the search process. Additional features include multi-point batch proposal, parallel execution as well as visualization and sophisticated logging mechanisms, which is especially useful for teaching and understanding of algorithm behavior. 'mlrMBO' is implemented in a modular fashion, such that single components can be easily replaced or adapted by the user for specific use cases.

Ratio-of-Uniforms Sampling for Bayesian Extreme Value Analysis

Provides functions for the Bayesian analysis of extreme value models. The 'rust' package < https://cran.r-project.org/package=rust> is used to simulate a random sample from the required posterior distribution. The functionality of 'revdbayes' is similar to the 'evdbayes' package < https://cran.r-project.org/package=evdbayes>, which uses Markov Chain Monte Carlo ('MCMC') methods for posterior simulation. In addition, there are functions for making inferences about the extremal index, using the models for threshold inter-exceedance times of Suveges and Davison (2010) and Holesovsky and Fusek (2020) . Also provided are d,p,q,r functions for the Generalised Extreme Value ('GEV') and Generalised Pareto ('GP') distributions that deal appropriately with cases where the shape parameter is very close to zero.