基于改进K-means聚类的欠定盲分离算法  

Underdetermined Blind Source Separation Based on Improved K-means Clustering

作　　者：柴文标[1] 何选森[1] 

机构地区：[1]湖南大学信息科学与工程学院,长沙410082

出　　处：《微计算机信息》2012年第10期391-392,453,共3页Control & Automation

摘　　要：对于K-means聚类算法,本文从数据预处理,区域划分法估计初始聚类中心等方面进行改进。本文算法改进了K-means算法对样本输入顺序及初始聚类中心很敏感的不足之处,并将改进后的算法对稀疏源混合信号数据进行聚类,从而估计出欠定盲分离的混合矩阵;在此基础上,利用最短路径法分离出源信号。仿真实验表明,相比于经典的稀疏源混合矩阵盲估计算法,本文算法具有更强的鲁棒性和更高的估计精度。For K-means clustering algorithm, this paper, mainly from the data pretreatment, Zone division method that estimates ini- tial clustering center areas to improve. The algorithm improves the deficiency of the K-means, which is sensitive to the sample input order and initial clustering center. The improved algorithm is applied to cluster the sparse source data of mixed signal, so as to esti- mate the mixed matrix of underdetermined blind source separation. And then, the shortest path algorithm is used to recover source signals.Experimental results show that the proposed mixing matrix recovery algorithm has advantages of high robustness and accuracy compared with conventional algorithms .

关 键 词：稀疏信号 K—means 中心点划分 

分 类 号：TN911[电子电信—通信与信息系统]