Title
|
* ! Graphical Methods in Bayesian Statistics
|
Date / Time / Room
|
Sponsor
|
Type
|
08/15/2002
8:30 AM -
10:20 AM
Room: H-Sutton Parlor Center
|
Section on Physical & Engineering Sciences*, Section on Statistical Graphics*
|
Invited
|
Organizer:
|
Robert E. McCulloch, University of Chicago
|
Chair:
|
Nicholas Polson, University of Chicago
|
Discussant:
|
|
Floor Discussion
|
10:15 AM
|
Description
Bayesian statistical approaches are particularly well suited to interactive, graphic representation. Ideally, the entire process of model/prior specification/elicitation, utility specification, and decision should be presented to the user in a comprehensible way that elucidates the impact of the various modeling assumptions. In high-dimensional models this is not always easily done. Sophisticated graphics and interactive user interfaces can help. Modern computing continues to advance rapidly. Papers in this session represent the ongoing effort to use computing technology to graphically represent Bayesian Analysis.
|