摘要
该文在盲分离开关算法的基础上提出一种峭度宽松开关算法。该算法用峭度作为激活函数中的开关量分析随机变量的高斯性,解决了原开关算法中衡量参数的不稳健性。该算法与盲分离开关算法和扩展的Infomax算法的仿真实验比较表明,新算法具有更好的分离效果和抗噪声能力。
This paper presents a new kurtosis loose switching algorithm （KLSA） based on the blind source separation switching algorithm. The algorithm directly uses kurtosis as a switching variable of the activation function to analyze the Gaussianity of random variables and it can solve the problem that the measurement parameter of the switch algorithm is not robust. Compared the performance of the new algorithm with the performance of BSS switching algorithm and extended Infomax ICA algorithm, the simulation results show that the new algorithm can get the better separation performance and anti - noise ability.
出处
《杭州电子科技大学学报（自然科学版）》
2007年第1期33-37,共5页
Journal of Hangzhou Dianzi University：Natural Sciences
关键词
盲分离
峭度
开关算法
blind source separation （BSS）
kurtosis
switching algorithm
