Abstract
This paper presents a new algorithm to solve the Independent Component Analysis (ICA) problem that has a very low computational complexity. The most remarkable feature of the proposed algorithm is that it does not need to compute higher-order statistics (HOS). In fact, the algorithm is based on trying to guess the sign of the independent components, after which it approximates the rest of the values. © 2011 Springer-Verlag Berlin Heidelberg.
Author supplied keywords
Cite
CITATION STYLE
Martín-Clemente, R., Hornillo-Mellado, S., & Camargo-Olivares, J. L. (2011). Independent component analysis: A low-complexity technique. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6687 LNCS, pp. 324–332). https://doi.org/10.1007/978-3-642-21326-7_35
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
