论文部分内容阅读
本文依据离散仿射小波变换理论,给出了小波神经网络模型的几种典型形式,详细讨论了小波基函数网络(WBFN)的时(空)—频域逼近原理,提出了该网络结构设计的“分解—综合”方法.仿真实验结果证实了上述理论和方法的有效性.
Based on the theory of discrete affine wavelet transform, several typical forms of wavelet neural network model are given. The principle of time-space-frequency approximation of wavelet basis function network (WBFN) is discussed in detail. The design of the network structure “Decomposition - comprehensive ” method.The simulation results verify the validity of the above theories and methods.