|
Activity Number:
|
504
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #307526 |
|
Title:
|
Capturing Uncertainty When the Event Probability Is Subject to Uncertainty
|
|
Author(s):
|
Ehsan S. Soofi*+ and Paul C. Nystrom and Masoud Yasai-Ardekani
|
|
Companies:
|
University of Wisconsin-Milwaukee and University of Wisconsin-Milwaukee and George Mason University
|
|
Address:
|
School of Business Administration, Milwaukee, WI, 53201,
|
|
Keywords:
|
maximum entropy ; prior ; posterior ; incomplete information
|
|
Abstract:
|
The problem of quantifying uncertainty when an event and its probability are under uncertainty arises in various fields of applications. We present various information theoretic solutions for the problem when the data include a point estimate and/or an interval estimate with or without a confidence level for the event probability. The information theoretic procedures provide measures that enable us to quantify and compare uncertainties for the intuitively clear cases in a precise manner. The entropy analysis also takes into account the trade-offs between confidence level, interval length, symmetry/asymmetry of the interval about the point estimate, and the closeness of the point estimate to certainty points 0 or 1, as well.
|
