Statistical and probabilistic fundamentals of ICA
Mark Girolami · 2004
Independent component analysis (ICA) can be viewed both as a multivariate statistical transformation and as describing a probabilistic generative latent variable model. Both of these viewpoints form the basis of a number of ICA algorithms and methods currently available. This talk will present the basic theoretical representations which underpin most ICA algorithms commonly available for various biomedical applications. (12 pages)