In this article we put forward a Bayesian approach for finding classif
ication and regression tree (CART) models. The two basic components of
this approach consist of prior specification and stochastic search. T
he basic idea is to have the prior induce a posterior distribution tha
t will guide the stochastic search toward more promising CART models.
As the search proceeds, such models can then be selected with a variet
y of criteria, such as posterior probability, marginal likelihood, res
idual sum of squares or misclassification rates. Examples are used to
illustrate the potential superiority of this approach over alternative
methods.