基于核独立成分分析的盲源信号分离  

Blind Source Separation Based on Kernel Independent Component Analysis

作　　者：姚伏天[1] 金连甫[1] 戴光[1] 

机构地区：[1]浙江大学计算机科学与工程学院,杭州310027

出　　处：《计算机工程与应用》2004年第6期44-46,共3页Computer Engineering and Applications

基　　金：浙江省自然科学基金重点项目(编号:ZD0223);浙江省自然科学基金(编号:601153)

摘　　要：独立成分分析(ICA)已经广泛用于盲源信号的分离(BSS)。论文介绍了基于核空间的ICA的原理和基本算法,然后介绍了该算法与典型ICA和主成分分析(PCA)在盲源信号分离中的比较。实验表明在盲源信号分离中,基于核空间的ICA与其他典型ICA和PCA算法相比更具有准确性和鲁棒性。Independent component analysis(ICA)is a main approach of blind source separation(BSS)and has received attention because of its potential application in signal processing.In this paper,kernel independent component analysis(KICA)'s principle and algorithm are introduced,and then the KICA comparison with some other ICA and principal component analysis(PCA)is given.The result of experiment shows that KICA is more accurate and robust than other conical ICA and PCA in the BBS.

关 键 词：独立成分分析(ICA) 核空间 盲源信号分离(BSS) 主成分分析(PCA) 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]