摘要
综述了独立分量分析(ICA)的基本原理及基于信息最大化原理的各种方法及其特性,介绍了HJ网络、基于信息最大化的Infomax法及其扩展算法、极大似然估计(MLE)法、负熵最大化法、基于高阶累量的ICA法和Bussage法,对各种方法性能做了比较,说明了ICA在生物医学信号处理中的应用,并对ICA的发展作了展望。
This paper summarizes the principles of independent component analysis (ICA) and various information maximum-based algorithms, including the HJ neural network, the Infomax and its extentions, the maximum likelihood approach, masimum negentropy, the high-order cumulant-based algorithm and bussage algorithm. The performance of the methods are compared, their applications in biomedical signal processing are presented, and the future development is indicated.
出处
《大连铁道学院学报》
2003年第2期64-69,共6页
Journal of Dalian Railway Institute
基金
国家自然科学基金(30170259)
辽宁省科学技术基金(2001101057)
