Bayesian Additive Regression Trees using Bayesian Model Averaging

Title:  Bayesian Additive Regression Trees using Bayesian Model Averaging

Speaker: Dr. Belinda Hernandez, School of Mathematics and Statistics, UCD.

Date: Thursday, 10th March 2016

Time: 4pm

Location: Room E0.01, Science East

Abstract:

Bayesian Additive Regression Trees (BART) is a statistical sum of trees model. It can be considered a Bayesian version of machine learning tree ensemble methods where the individual trees are the base learners. However for datasets where the number of variables is large the algorithm can become very computationally expensive. This is mainly due to its method for searching and proposing new split rules while growing trees being inefficient. We propose an alternative algorithm for BART called BART-BMA which uses Bayesian Model Averaging and a greedy search algorithm to produce a model which is much more efficient than BART for high dimensional data.

Series: Statistics