正交泛函网络函数逼近理论及算法

基于正交函数的概念和特性,提出一种正交泛函网络新模型,给出了正交泛函网络学习算法.该算法是借助于正交函数性质和Lagrange乘数法做辅助函数,对泛函参数学习过程归结为求解一组线性方程组的过程.最后,通过函数逼近算例计算机仿真结果表明,该算法十分有效,具有模型简单、逼近精度高等特点.

Orthogonal Function Network Approximate Theory and Learning Algorithm

This paper presented a new kind of functional network model which is based on orthogonal functions,a learning algorithm of orthogonal functional network was proposed,the learning of function parameters uses orthogonal function characteristic and Lagrange multipliers by means of auxiliary function and solving a system of linear equation obtains functions parameters.Finally,An experiment in approximating typical functions was given.The simulation results show that the learning algorithm presented in the paper...