The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Abstract Details
Activity Number:
|
119
|
Type:
|
Topic Contributed
|
Date/Time:
|
Monday, August 1, 2011 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Bayesian Statistical Science
|
Abstract - #301285 |
Title:
|
Issues in Bayesian Datamining
|
Author(s):
|
John Liechty*+
|
Companies:
|
Penn State University
|
Address:
|
409 BB, Univeristy Park, PA, 16802,
|
Keywords:
|
Markov chain Monte Carlo ;
Data Mining ;
Parallel Computing
|
Abstract:
|
One advantage of using the Markov chain Monte Carlo (MCMC) algorithm for inference of hierarchical Bayesian models on extreme size data sets is that the sampling algorithm and data can be distributed to multiple nodes of a high performance computer in such a way that draws from full conditional distributions can be accommodated by passing just summaries of relevant parameters, between the different nodes as opposed to passing the data between nodes. These modeling and computation approaches allow for complex inference on extreme data sets in an efficient manner - allowing for schemes that analyze the entire data set as opposed to random subsets. In addition to providing some illustrative examples and results of computer experiments, we discuss the theoretical, convergence properties of the parallelized MCMC algorithm when using an asynchronous scheme for sampling and sharing parameter values across multiple nodes.
|
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 2011 program
|
2011 JSM Online Program Home
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.