Abstract
Expectation-Maximization (EM) algorithms for independent component analysis are presented in this paper. For super-Gaussian sources, a variational method is employed to develop an EM algorithm in closed form for learning the mixing matrix and inferring the independent components. For sub-Gaussian sources, a symmetrical form of the Pearson mixture model (Neural Comput. 11 (2) (1999) 417-441) is used as the prior, which also enables the development of an EM algorithm in fclosed form for parameter estimation.
Original language | English |
---|---|
Pages (from-to) | 503-512 |
Number of pages | 10 |
Journal | Neurocomputing |
Volume | 61 |
Issue number | 1-4 |
DOIs | |
State | Published - Oct 2004 |
Keywords
- EM algorithm
- Independent component analysis
- Overcomplete representations
- Variational method