PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

An accelerated algorithm for MCMC Bayesian decision tree sampling
Gilles Blanchard
In: Workshop Ensemble Methods, 4-5 march 2005.

Abstract

We present a novel type of MCMC sampler intended to deal with a Bayesian framework for classification and regression trees. Typically, it is hard in practice to sample from the Bayesian posterior on trees (in order to approximate posterior-averaged or MAP estimators) by naive MCMC methods, because the space of trees is huge and the posterior presents a lot of local minima. The algorithm we introduce allows, for a large class of prior distributions, to develop an accelerated MCMC algorithm. The key point is the exact computation of some conditional posterior averages over tree structures at each step. This results in a dramatic increase of the number of models actually taken into account. It can be used in a variety of situations for Bayesian inference and we show its practical relevance on a demonstration example.

PDF - PASCAL Members only - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Conference or Workshop Item (Invited Talk)
Additional Information:These are only the slides of the talk. (A related paper, in french, was published in 2004 -- but was not submitted in the PASCAL EPrints because it was not in english)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Computational, Information-Theoretic Learning with Statistics
Learning/Statistics & Optimisation
Theory & Algorithms
ID Code:1351
Deposited By:Gilles Blanchard
Deposited On:28 November 2005