Bayesian CART model search

Hugh A. Chipman, Edward I. George, Robert McCulloch

Research output: Contribution to journalArticle

336 Citations (Scopus)

Abstract

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

Original languageEnglish (US)
Pages (from-to)935-948
Number of pages14
JournalJournal of the American Statistical Association
Volume93
Issue number443
StatePublished - Sep 1998
Externally publishedYes

Fingerprint

Classification and Regression Trees
Bayesian Classification
Stochastic Search
Misclassification Rate
Marginal Likelihood
Posterior Probability
Sum of squares
Posterior distribution
Bayesian Approach
Model
Specification
Alternatives
Classification and regression trees

Keywords

  • Binary trees
  • Markov chain Monte Carlo
  • Mixture models
  • Model selection
  • Model uncertainty
  • Stochastic search

ASJC Scopus subject areas

  • Mathematics(all)
  • Statistics and Probability

Cite this

Chipman, H. A., George, E. I., & McCulloch, R. (1998). Bayesian CART model search. Journal of the American Statistical Association, 93(443), 935-948.

Bayesian CART model search. / Chipman, Hugh A.; George, Edward I.; McCulloch, Robert.

In: Journal of the American Statistical Association, Vol. 93, No. 443, 09.1998, p. 935-948.

Research output: Contribution to journalArticle

Chipman, HA, George, EI & McCulloch, R 1998, 'Bayesian CART model search', Journal of the American Statistical Association, vol. 93, no. 443, pp. 935-948.
Chipman, Hugh A. ; George, Edward I. ; McCulloch, Robert. / Bayesian CART model search. In: Journal of the American Statistical Association. 1998 ; Vol. 93, No. 443. pp. 935-948.
@article{e07072a6e830435b9f4bf145ab680fea,
title = "Bayesian CART model search",
abstract = "In this article we put forward a Bayesian approach for finding classification and regression tree (CART) models. The two basic components of this approach consist of prior specification and stochastic search. The basic idea is to have the prior induce a posterior distribution that will guide the stochastic search toward more promising CART models. As the search proceeds, such models can then be selected with a variety of criteria, such as posterior probability, marginal likelihood, residual sum of squares or misclassification rates. Examples are used to illustrate the potential superiority of this approach over alternative methods.",
keywords = "Binary trees, Markov chain Monte Carlo, Mixture models, Model selection, Model uncertainty, Stochastic search",
author = "Chipman, {Hugh A.} and George, {Edward I.} and Robert McCulloch",
year = "1998",
month = "9",
language = "English (US)",
volume = "93",
pages = "935--948",
journal = "Journal of the American Statistical Association",
issn = "0162-1459",
publisher = "Taylor and Francis Ltd.",
number = "443",

}

TY - JOUR

T1 - Bayesian CART model search

AU - Chipman, Hugh A.

AU - George, Edward I.

AU - McCulloch, Robert

PY - 1998/9

Y1 - 1998/9

N2 - In this article we put forward a Bayesian approach for finding classification and regression tree (CART) models. The two basic components of this approach consist of prior specification and stochastic search. The basic idea is to have the prior induce a posterior distribution that will guide the stochastic search toward more promising CART models. As the search proceeds, such models can then be selected with a variety of criteria, such as posterior probability, marginal likelihood, residual sum of squares or misclassification rates. Examples are used to illustrate the potential superiority of this approach over alternative methods.

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

KW - Binary trees

KW - Markov chain Monte Carlo

KW - Mixture models

KW - Model selection

KW - Model uncertainty

KW - Stochastic search

UR - http://www.scopus.com/inward/record.url?scp=0032346848&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0032346848&partnerID=8YFLogxK

M3 - Article

VL - 93

SP - 935

EP - 948

JO - Journal of the American Statistical Association

JF - Journal of the American Statistical Association

SN - 0162-1459

IS - 443

ER -