机构地区: 广东商学院
出 处: 《中山大学学报论丛》 1996年第5期176-180,共5页
摘 要: 小波分析已成为众多领域中的有力工具.本文采用调制的小波基对输入模式预处理,在函数链神经网络的基础上设计了小波型神经网络WTNN,从而把小波变换与函数链神经网络有机地结合起来.这样设计的模型有惊人的学习速度;体系结构的通用性好;适应性强等特点.最后。 Wavelets analysis have become powerful tools in lots of fields. This paper takes modulation wave lets coordinate to pre transform the learning specimen. On the basis of the functional link nevral network, we design the wavelets neural network model WTNN and combine wavelets transform with the functional link neural network successfully. This new model has distinguishing feature as following:(1) surprising learning speed; (2) commonly used structure in erea of neural networks; (3) strong adaptability. Lastly, we give some examples to prove WTNN model has distinguishing learning speed.