AN IMPROVED KICA ALGORITHM BASED ON KECA

摘要：

Kernel Independent Component Analysis (KICA) is a non-linear method for blind source separation (BSS) advanced recently. It has a high computational complexity. Recurring to the characteristic of dimension reduction and redundancy removing of kernel entropy component analysis (K.ECA), an improved algorithm of KICA, kernel entropy independent component analysis (KEICA), is proposed to reduce the high computational complexity of KICA. In a simulation experiment with the separation performance, Amari error and the accumulated time as the criterions, KEICA is compared to KICA. Results show that KEICA is superior to KICA comprehensively.

关键词： Kernel independent component analysis Kernel entropy component analysis Dimension reduction

作者: XU YANG XUE-YUAN ZHANG BAO-QING LI

作者单位: Shanghai Institute of Microsystem and Information technology of Chinese Academy of Science Shanghai, China

会议类型: 国际会议

会议名称: 3rd International Conference on Mechanical and Electrical Technology(ICMET2011) (2011第三届机械与电气技术国际会议)

会议地点: 大连

会议语种:英文

页码: 513-517

在线出版日期: 2011-08-26（万方平台首次上网日期，不代表论文的发表时间）