二维随机变量独立性度量及其在独立分量分析中的应用  

Measuring dependence of bivariate signals and its extensions in independent component analysis

作　　者：索文莉[1] 李长国[2] 樊顺厚[1] 

机构地区：[1]天津工业大学理学院,天津300160 [2]解放军军事交通学院基础部,天津300161

出　　处：《天津工业大学学报》2007年第6期82-85,共4页Journal of Tiangong University

摘　　要：以调整的2χ统计量作为随机信号相关性的度量,以该度量作为独立分量分析的对照函数,利用模拟退火算法分离出了参数办法不容易分离出来的独立信号.An adjusted Chi-Square Statistic is proposed as the measure of dependence of random signals, and this measure is taken as the contrast function of ICA. Taking simulated annealing method the random signal can be separated which can not be separated by parametric method.

关 键 词：独立分量分析 相关性 模拟退火 

分 类 号：O211.5[理学—概率论与数理统计]