Loading...

Using multivariate score functions in source separation: Application to post non-linear mixtures

Babaie Zadeh, M ; Sharif University of Technology | 2002

3 Viewed
  1. Type of Document: Article
  2. Publisher: Sharif University of Technology , 2002
  3. Abstract:
  4. In this paper, Joint Score Function (JSF) and Marginal Score Function (MSF) are first defined. It is then pointed out that their difference (SFD) can be treated as the stochastic gradient of mutual information and, hence, can be used in minimizing the mutual information with gradient-based methods. An estimator for SFD is then presented, based on nonlinear regression by means of spline smoothing. It is shown that SFD can be used to obtain a new non parametric algorithm for source separation in Post Non-Linear (PNL) mixtures. The method is very general and can be extended to convolutive mixtures, which is currently being studied
  5. Keywords:
  7. Source: Scientia Iranica ; Volume 9, Issue 4 , 2002 , Pages 409-418 ; 10263098 (ISSN)
  8. URL: http://scientiairanica.sharif.edu/article_2690.html