基于小波变换的全局最优独立分量分离算法  

ICA method with global optimal property based on wavelet transformation

作　　者：毕雪[1] 陈向东[2] 李湃[2] 

机构地区：[1]西华大学网络中心,四川成都610039 [2]西南交通大学信息科学与技术学院,四川成都610031

出　　处：《传感器与微系统》2007年第8期102-104,107,共4页Transducer and Microsystem Technologies

基　　金：国家自然科学基金资助项目(60572026)

摘　　要：提出一种利用小波变换的低频系数作为滑动因子的盲源分离算法,通过建立以分母作为预测误差的信噪比目标函数,优化过程导致广义特征值求解,广义特征值构成的特征向量就是要求的分离矩阵,求解过程无需迭代。通过实验证明:计算量小,分离精度高,相似系数分别到达了1和0.9998,同时,可根据需要灵活选择小波基。A kind of independent component analysis （ICA） method with global optimal property is proposed, which is based on low frequency part of wavelet transformation as moving factors and built a signal-to-noise ratio objective function on predictability error of denominator. Optimizing leads to get generalized eigenvalue and eigenvector which is the demanding separable matrix. Iteration is not necessary. The experiments show that computation is small and separable precision is higher,which can respectively get resembled coefficients of 1 and O. 999 8. Meanwhile it can choose wavelet basis flexibly based on different requirement.

关 键 词：独立分量分析 全局最优 小波变换 滑动 逼近 

分 类 号：TN911.6[电子电信—通信与信息系统]