Indexed by:Journal Article

Date of Publication:2012-01-01

Journal:ICIC Express Letters

Included Journals:Scopus、EI

Volume:6

Issue:9

Page Number:2357-2362

ISSN:1881803X

Abstract:A novel Independent Component Analysis algorithm based on parameterized model (PMICA) is proposed which can separate the mixture of super- and sub-Gaussian sources. Two approximate probability density function models are proposed respectively. In the framework of natural gradient, the parameters of two models are adaptively manipulated by online kurtosis learning. Stability analysis is discussed which provides the ranges of models parameters. Experiments show that the proposed PMICA algorithm has better performance than some other algorithms. ? 2012 ICIC International.