节点文献
函数联接神经网络的实验研究
Experimental Research on Functional Link Neural Network
【摘要】 函数联接神经网络具有结构简单 ,易于实现 ,扩展性和通用性强等许多优点 .利用面向对象方法实现了函数联接神经网络的几种模型 ,探讨了该网络的变结构调整问题 ,并通过实际的例子对网络的不同模型进行了实验研究 ,提出了一个新的学习因子动态调整算法
【Abstract】 Numerous and significant advantages accrue from using a functional link network(FLN),including simplification in architecture as well as computational procedures,and uniform architecture for extension and generalization.In this paper,several models of the FLN and their implementation for dynamic structure are analyzed by using the object oriented method.Some experiments for the different models have been done by a useful application,and a new algoriothm for dynamically adjusting the learning parameter is proposed and implemented.
【关键词】 函数联接网络;
在线学习;
离线学习;
阻尼振荡学习因子;
【Key words】 functional-link network; online learning; offline learning; damped oscillation learning pa-rameter;
【Key words】 functional-link network; online learning; offline learning; damped oscillation learning pa-rameter;
- 【文献出处】 长沙电力学院学报(自然科学版) ,JOURNAL OF CHANGSHA UNIVERSITY OF ELECTRIC POWER(NATURAL SCIENCE) , 编辑部邮箱 ,2000年02期
- 【分类号】TP18
- 【被引频次】3
- 【下载频次】41