Efficient Independent Component Analysis (ii)

Abstract

Independent component analysis (ICA) has been widely used in separating hidden sources from observed linear mixtures in many fields such as brain imaging analysis, signal processing, telecommunication. Many statistical techniques based on M-estimates have been proposed in estimating the mixing matrix. Recently a few methods based on nonparametric tools are also available. However, in-dept analysis on the convergence rate and asymptotic efficiency has not been available. In this paper we analyze ICA under the framework of semiparametric theories [see Bickel, Klaassen, Ritov and Wellner (1993)] and propose a straightforword estimate based on the efficient score function by using B-spline approximations. This estimate exhibits better performance than stardard ICA methods in a variety of simulations. It is proved that this estimate is Fisher efficient under moderate conditions.

Extracted Key Phrases

2 Figures and Tables

Cite this paper

@inproceedings{ChenEfficientIC, title={Efficient Independent Component Analysis (ii)}, author={Aiyou Chen and Peter J. Bickel} }