Title: Average derivative estimation with bayesian decision tree ensembles
Authors: Christoph Breunig - Emory University (United States) [presenting]
Abstract: The estimation of average derivatives with Bayesian decision tree ensembles is considered. We make use of soft decision trees in which the decisions are treated as probabilistic. Specifically, we make use of the Bayesian additive regression trees framework. The posterior distribution concentrates at the minimax rate for estimating directional derivatives (up to a logarithmic factor) for sparse functions and functions with additive structures in the high dimensional regime. We illustrate the finite sample properties in simulations and empirical illustrations.