BART (Bayesian Additive Regression Tree) is a nonparametric Bayesian regression approach which enables full posterior inference including point and interval estimates of unknown regression function as well as the marginal effects of potential predictors.