基于MATLAB优化工具箱函数的ICA求解  

MATLAB-based ICA solution of optimized toolbox function

作　　者：刘颖[1] 严军[1] 

机构地区：[1]西北师范大学化学化工学院,甘肃兰州730070

出　　处：《兰州理工大学学报》2014年第2期98-101,共4页Journal of Lanzhou University of Technology

基　　金：甘肃省自然科学基金(1107RJZA136)

摘　　要：在系统分析FastICA法的基础上,给出基于牛顿法迭代格式的计算过程和计算步骤.应用MATLAB优化工具箱中成熟的算法,以平方阶的随机变量信息熵作为目标函数对FastICA算法进行改进,使得计算更稳定.实验表明,应用MATLAB优化工具箱函数的求解方法可行、实用,对ICA在工程实际中的推广和应用具有积极作用.On the basis of the FastlCA method for systematic analysis , the calculation process and calculation procedure of Newton method-based iterative format was given. The algorithm in FastlCA was improved with the mature algorithm in MATLAB optimized toolbox and taking the square-ordered random variable information entropy as object function. Experiment showed that the solution was feasible and practical, having an active function for generalization and application of ICA in engineering practice.

关 键 词：独立分量分析 FASTICA 优化 MATLAB 

分 类 号：TP14[自动化与计算机技术—控制理论与控制工程]