Super-fast BART (Bayesian Additive Regression Trees) in Python
Likelihood types for each outcome component in the regression.
Continuous outcome with Normal error.
Binary outcome in {0, 1} with probit link.