531f10GIBBS1

# 531f10GIBBS1 - STAT 531: Gibbs Sampler HM Kim Department of...

This preview shows pages 1–7. Sign up to view the full content.

STAT 531: Gibbs Sampler HM Kim Department of Mathematics and Statistics University of Calgary Fall 2010 1/23

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
The Gibbs sampler was introduced in the context of image processing by Geman and Geman, 1984 It is a special case of MH samplings The task remains to specify how to construct a Markov chain whose values converge to the target distribution. Can we know the joint distribution simply by knowing the full conditional distributions? Fall 2010 2/23
Hammersley-Cliﬀord Theorem (for two blocks) Suppose we have a joint density f ( x , y ). The theorem proves that we can write out the joint density in terms of the conditional densities f ( x | y ) and f ( y | x ): f ( x , y ) = f ( y | x ) R f ( y | x ) f ( x | y ) dy We can write the denominator as Z f ( y | x ) f ( x | y ) dy = Z f ( x , y ) f ( x ) f ( x , y ) f ( y ) dy = Z f ( y ) f ( x ) dy = 1 f ( x ) Fall 2010 3/23

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Then f ( x , y ) = f ( y | x ) 1 f ( x ) = f ( x | y ) f ( x ) = f ( x , y ) The theorem shows that knowledge of the conditional densities allows us to get the joint density. This works for more than two blocks of parameters. Fall 2010 4/23
The key to the Gibbs sampler is that one only considers univariate conditional distributions . Such conditional distributions are far easier to simulate than complex joint distributions and usually have simple forms. Thus one simulates q random variables sequentially from the q univariate conditionals rather than generating a single q -dimensional vector in a single pass using the full joint distributions. Fall 2010 5/23

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
, Example : To introduce the Gibbs sampler, consider a bivariate random variable ( x , y ), and suppose we wish to compute one or both marginals, p ( x ) and p ( y ). The idea behind the sampler is that it is far easier to consider a
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 02/04/2011 for the course STAT 531 taught by Professor Gaborlukacs during the Spring '11 term at Manitoba.

### Page1 / 23

531f10GIBBS1 - STAT 531: Gibbs Sampler HM Kim Department of...

This preview shows document pages 1 - 7. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online