CAJ | 학술논문

对于传统的神经网络中神经元模型在结构和信息存储能力上存在的不足,本文提出了一种基于广义小波基函数网络的神经元集聚模型.这种小波神经网络不仅收敛速度快,非线性逼近能力更好,而且具有内部结构变尺度、自适应调整和广义信息存储等智能化特点,更符合生物原型的实际情况.静态学习和准动态学习仿真实验证明这种神经网络结构的有效性.
대우전통적신경망락중신경원모형재결구화신식존저능력상존재적불족,본문제출료일충기우엄의소파기함수망락적신경원집취모형.저충소파신경망락불부수렴속도쾌,비선성핍근능력경호,이차구유내부결구변척도、자괄응조정화엄의신식존저등지능화특점,경부합생물원형적실제정황.정태학습화준동태학습방진실험증명저충신경망락결구적유효성.
Considering the deficiency in the structure as well as in t he capability of information storage of the neuron models used in traditional ne uron networks,an intelligent neuron assemblage model with generalized wavelet ba sis function network as its stimulating function is proposed.The wavelet neural network not only converges much faster with better nonlinear approaching capabil ity,but also is characterized by its ability in variable-scale adaptive structu re adjustment and the generalized information storage,etc,which guarantee the ne twork reflects the biological prototype much more faithfully.Static and pseudo d ynamic learning is demonstrated to prove the validity of the proposed mechanism.