bartMachine: Machine Learning with Bayesian Additive Regression Trees

TL;DR

The paper introduces bartMachine, an R package for Bayesian additive regression trees that enhances data analysis with features like variable selection, interaction detection, and improved speed, scalability, and usability.

Contribution

It presents a new, faster, and feature-rich R package for BART, including parallelization and handling of large, high-dimensional datasets.

Findings

Significantly faster than existing R implementations

Supports variable selection and interaction detection

Capable of handling large samples and high-dimensional data

Abstract

We present a new package in R implementing Bayesian additive regression trees (BART). The package introduces many new features for data analysis using BART such as variable selection, interaction detection, model diagnostic plots, incorporation of missing data and the ability to save trees for future prediction. It is significantly faster than the current R implementation, parallelized, and capable of handling both large sample sizes and high-dimensional data.