JSM 2004 - Toronto

Abstract #301697

This is the preliminary program for the 2004 Joint Statistical Meetings in Toronto, Canada. Currently included in this program is the "technical" program, schedule of invited, topic contributed, regular contributed and poster sessions; Continuing Education courses (August 7-10, 2004); and Committee and Business Meetings. This on-line program will be updated frequently to reflect the most current revisions.

To View the Program:
You may choose to view all activities of the program or just parts of it at any one time. All activities are arranged by date and time.

The views expressed here are those of the individual authors
and not necessarily those of the ASA or its board, officers, or staff.


Back to main JSM 2004 Program page



Activity Number: 106
Type: Topic Contributed
Date/Time: Monday, August 9, 2004 : 10:30 AM to 12:20 PM
Sponsor: Section on Bayesian Statistical Science
Abstract - #301697
Title: Intuitive, Easily Programmed Steps that Reduce MCMC Autocorrelation
Author(s): Todd L. Graves*+ and Paul L. Speckman and Dongchu Sun
Companies: Los Alamos National Laboratory and University of Missouri, Columbia and University of Missouri, Columbia
Address: MS F600, PO Box 1663, Los Alamos, NM, 87545,
Keywords: Metropolis algorithm ; YADAS ; Bayesian computation
Abstract:

Often, an analyst writes software to analyze a Bayesian model using Markov chain Monte Carlo, only to find that high correlations between parameters cause mixing to be so poor as to render the algorithm useless. We will demonstrate how augmenting existing algorithms with Metropolis(-Hastings) steps in well-chosen directions can dramatically improve mixing. These directions can usually be found through easy analysis of individual terms in the posterior distribution. The MCMC software package YADAS makes it particularly convenient to create algorithms featuring these steps.


  • The address information is for the authors that have a + after their name.
  • Authors who are presenting talks have a * after their name.

Back to the full JSM 2004 program

JSM 2004 For information, contact jsm@amstat.org or phone (888) 231-3473. If you have questions about the Continuing Education program, please contact the Education Department.
Revised March 2004