|
Activity Number:
|
509
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #309442 |
|
Title:
|
A New Approach to the Inverse Bayes Formula of Compatible Conditional Distributions
|
|
Author(s):
|
Kun-Lin Kuo*+ and Chwan-Chin Song and Thomas J. Jiang
|
|
Companies:
|
National Chengchi University and National Chengchi University and National Chengchi University
|
|
Address:
|
Dept. of Math, 64 ChiNan Road Section 2, Wen-Shan, Taipei, 11605, Taiwan
|
|
Keywords:
|
compatibility ; joint density ; conditional density ; marginal density ; rank one positive extension
|
|
Abstract:
|
The inverse Bayes formula (IBF) is a statistical tool which is important in distribution theory and Bayesian missing data problems (examples can be seen in Tian and Tan (2003), among others). Ng (1997) provided a form of IBF essentially for product measurable space and recognized that IBF is potentially useful in computing marginals and checking compatibility. Recently, Tian and Tan (2003) provided a form of modified IBF in nonproduct measurable space and gave some applications. Although the modified IBF is useful, it holds only in a restricted space. In this article, we give a new theoretical approach to the IBF that can be applied in more general space and in both discrete and continuous cases. This new approach provides the rationale of Ng (1997)'s IBF and may also solve some problems that Tian and Tan (2003) can't do. Examples are given to illustrate our new approach to the IBF.
|
