作 者: ;
机构地区: 东南大学信息科学与工程学院无线电工程系
出 处: 《模式识别与人工智能》 2004年第2期178-183,共6页
摘 要: 提出了一种新颖的有导-无导联合学习神经网络盲源分离方法.该方法用经典的有导学习误差反传规则训练多层感知器,对信号的概率分布函数进行自适应逼近,从而给出分离矩阵无导学习规则中非线性激励函数的恰当估计.将该方法与纯粹无导学习方法进行了对比,说明了其性能上的优越性. A novel supervised- unsupervised combined neural learning approach for blind source separation is proposed. The classical supervised learning back-propagation rule is used to train multilayer perceptron for approximating the signal distribution functions adaptively, giving an appropriate estimate of the nonlinear activation function in the unsupervised learning rule of the de-mixing matrix. A comparison with purely unsupervised learning shows its superior performance.