|
Activity Number:
|
560
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Social Statistics Section
|
| Abstract - #304640 |
|
Title:
|
A Choice of the Number of Factors and Hyper-Parameter Selection in Bayesian Factor Analysis Model
|
|
Author(s):
|
Kei Hirose*+ and Shuichi Kawano and Sadanori Konishi and Masanori Ichikawa
|
|
Companies:
|
Kyushu University and Kyushu University and Kyushu University and Tokyo University of Foreign Studies
|
|
Address:
|
6-10-1 Hakozaki, Higashi-ku, Fukuoka, 812-8581, Japan
|
|
Keywords:
|
Factor analysis ; Bayesian approach ; EM algorithms ; Model selection criterion ; Improper solutions ; Number of factors
|
|
Abstract:
|
Factor analysis has been widely used in the social and natural sciences to explore the covariance structure among a set of observed random variables by constructing a smaller number of unobserved random variables called common factors. In maximum likelihood factor analysis, the estimates of unique variances can often turn out to be zero or negative. In order to overcome this difficulty, we use a Bayesian approach by specifying a prior distribution for unique variances. Crucial issues in Bayesian factor analysis model are the choice of hyper-parameters for the prior distribution and also the number of factors. We propose a Bayesian factor analysis modeling procedure that prevents the occurrence of improper solutions and also chooses the appropriate number of factors objectively.
|
