基于仿生模式识别的构造型神经网络分类方法.pdfVIP

第26卷第 9期 计 算 机 应 用 研 究 Vo.l 26 No. 9 2009年9月 ApplicationResearch of Computers Sep. 2009 * , (华南师范大学 计算机学院, 广州 510631) : 从 BP神经元模型和 RBF神经元模型几何意义出发, 将仿生模式识别理论引入到神经网络分类中, 提 出了一种基于仿生模式识别的构造型神经网络分类方法, 通过构造不同结构神经元结合的神经网络, 实现了对 不同类样本在高维空间中形成的不同形状几何体的覆盖实验证明该算法是非常有效的 : 仿生模式识别; 神经元; 构造型神经网络; 高维空间; 分类 : P391 : A : 2009) doi: 10. 3969 /j. issn. 2009. 09. 032 Constructive neuron networks classification algorithm based on bionic pattern recognition CHEN Q im a,i ZHOU H aiq ing (C ol lege of C omp u ter Sc ience, S outh Ch ina N orma l Un iversity, Guangzhou 510631, Ch ina) Ab stract: his paper introduced the theory of bionic pattern recognition into neuralnetworks classification, Firstly from geo metric significance of theBP neuronmodeland the RBF neuronmodel, proposed a new constructive neuron newt orks classifi cation algorithm based on bionic pattern recognition. hen covered high dmi ensional geometricaldistribution of the sample set in the feature space by constructing a new type ofANN. Finally, the test proves that the algorithm is very effective. K ey words: bionic pattern recognition; neuron; constructive neural newt orks; high dmi ensional; classification , 0 , [ 1] , , 2002 , , , , , 1 , 11 , , : , , , ,