节点文献
单调RBF神经网络的逼近性分析
Constructing Monotonic RBF Neural Networks to Approximate Monotonic Functions
【摘要】 针对工程中输入输出呈单调关系系统首先提出单调径向基神经网络,然后给出单调性条件定理,并证明用单调径向基神经网络插值可以逼近紧致集上任意单输入单输出的单调函数。
【Abstract】 In this paper, novel monotonic RBF neural networks are given first, then the conditions for SISO and MISO are developed. a constructive method is developed to establish the fact that we can build a monotonic RBF neural networks to (approximate any continous monotonic function on a compact set.)
【关键词】 单调径向基神经网络;
构造理论;
单调径向基神经网络插值;
【Key words】 Monotonic RBF Neural Networks; Constructive Theory; Monotonic RBF Neural Networks Interpolation;
【Key words】 Monotonic RBF Neural Networks; Constructive Theory; Monotonic RBF Neural Networks Interpolation;
【基金】 国家自然科学基金资助项目(60174021)
- 【文献出处】 系统工程 ,Systems Engineering , 编辑部邮箱 ,2004年08期
- 【分类号】TP273
- 【被引频次】4
- 【下载频次】181