Markov Chain Monte Carlo using Tree-Based Priors on Model Structure

Nicos Angelopoulos; James Cussens

Markov Chain Monte Carlo using Tree-Based Priors on Model Structure

Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

We present a general framework for defining priors on model structure and sampling from the posterior using the Metropolis-Hastings algorithm. The key ideas are that structure priors are defined via a probability tree and that the proposal distribution for the Metropolis-Hastings algorithm is defined using the prior, thereby defining a cheaply computable acceptance probability. We have applied this approach to Bayesian net structure learning using a number of priors and proposal distributions. Our results show that these must be chosen appropriately for this approach to be successful.

Original language	English
Title of host publication	Proceedings of the Seventeenth Annual Conference on Uncertainty in Artificial Intelligence (UAI--2001)
Editors	Jack Breese, Daphne Koller
Place of Publication	Seattle
Publisher	MORGAN KAUFMANN PUB INC
Pages	16-23
Number of pages	8
Publication status	Published - 1 Aug 2001

Access to Document

ftp://ftp.cs.york.ac.uk/pub/aig/Papers/james.cussens/uai01.ps.gz

Cite this

@inproceedings{e4cb88aef02e498185d3256b0977ba3f,

title = "Markov Chain Monte Carlo using Tree-Based Priors on Model Structure",

abstract = "We present a general framework for defining priors on model structure and sampling from the posterior using the Metropolis-Hastings algorithm. The key ideas are that structure priors are defined via a probability tree and that the proposal distribution for the Metropolis-Hastings algorithm is defined using the prior, thereby defining a cheaply computable acceptance probability. We have applied this approach to Bayesian net structure learning using a number of priors and proposal distributions. Our results show that these must be chosen appropriately for this approach to be successful.",

author = "Nicos Angelopoulos and James Cussens",

year = "2001",

month = aug,

day = "1",

language = "English",

pages = "16--23",

editor = "Jack Breese and Daphne Koller",

booktitle = "Proceedings of the Seventeenth Annual Conference on Uncertainty in Artificial Intelligence (UAI--2001)",

publisher = "MORGAN KAUFMANN PUB INC",

}

Markov Chain Monte Carlo using Tree-Based Priors on Model Structure. / Angelopoulos, Nicos ; Cussens, James.
Proceedings of the Seventeenth Annual Conference on Uncertainty in Artificial Intelligence (UAI--2001). ed. / Jack Breese; Daphne Koller. Seattle: MORGAN KAUFMANN PUB INC, 2001. p. 16-23.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Markov Chain Monte Carlo using Tree-Based Priors on Model Structure

AU - Angelopoulos, Nicos

AU - Cussens, James

PY - 2001/8/1

Y1 - 2001/8/1

N2 - We present a general framework for defining priors on model structure and sampling from the posterior using the Metropolis-Hastings algorithm. The key ideas are that structure priors are defined via a probability tree and that the proposal distribution for the Metropolis-Hastings algorithm is defined using the prior, thereby defining a cheaply computable acceptance probability. We have applied this approach to Bayesian net structure learning using a number of priors and proposal distributions. Our results show that these must be chosen appropriately for this approach to be successful.

AB - We present a general framework for defining priors on model structure and sampling from the posterior using the Metropolis-Hastings algorithm. The key ideas are that structure priors are defined via a probability tree and that the proposal distribution for the Metropolis-Hastings algorithm is defined using the prior, thereby defining a cheaply computable acceptance probability. We have applied this approach to Bayesian net structure learning using a number of priors and proposal distributions. Our results show that these must be chosen appropriately for this approach to be successful.

M3 - Conference contribution

SP - 16

EP - 23

BT - Proceedings of the Seventeenth Annual Conference on Uncertainty in Artificial Intelligence (UAI--2001)

A2 - Breese, Jack

A2 - Koller, Daphne

PB - MORGAN KAUFMANN PUB INC

CY - Seattle

ER -