|
Activity Number:
|
548
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Thursday, August 6, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #303799 |
|
Title:
|
Bayesian Independent Component Analysis Using Mixture Priors
|
|
Author(s):
|
Ani Eloyan*+ and Sujit Ghosh
|
|
Companies:
|
North Carolina State University and North Carolina State University
|
|
Address:
|
Department of Statistics, Raleigh, NC, 27695-8203,
|
|
Keywords:
|
Independent Component Analysis ; Mixture Prior ; Signal Separation
|
|
Abstract:
|
Independent Component Analysis (ICA) is a data manipulation technique commonly used by experimenters working in such research areas as statistics, neural networks, etc. ICA methods are widely implemented to find good representations for large data sets. In signal processing ICA is used to recover source signals using observed mixtures of the sources. There are a number of algorithms for solving this problem such as the commonly used FastICA method. Since in many applications we have a priori knowledge about the structure of the data Bayesian approach to the problem can be more efficient than the conventional methods. A Bayesian model with mixtures of normal densities as priors for the original source signals is proposed. The method is compared with the commonly used ICA algorithms using simulated data. The performance of the method is shown by real data applications.
|
