摘要
为满足工程应用中对数据处理的需要,讨论基于Matlab实现函数逼近的三种方法:插值、拟合和神经网络逼近。在介绍基本原理的基础上,利用Matlab的插值和拟合函数,结合实例对分段线性插值、Hermite、三次样条插值及最小二乘曲线拟合法的Matlab实现方法进行研究。设计非线性函数逼近的BP神经网络,通过网络训练、仿真达到了预期的效果。所有结果表明,采用不同的逼近方法,利用Matlab编程可以简单、有效地实现函数逼近。
Three means to realize function approach such as the interpolation approach,fitting approach as well as the neural network approach are discussed based on Matlab to meet the demand of data processing in engineering application.Based on basic principle of the introduction,realization methods to piecewise linear interpolation,Hermite interpolation,cubic spline interpolation and least squares curve fitting method are researched using interpolation function and fitting function in Matlab with example.BP neural network to proximate nonlinear functions is designed,desired elect is achieved through the training and simulation of network. All results indicate that different methods are adopted according to specific problems,and function approach is realized simply and effectively in Matlab.
出处
《现代电子技术》
2009年第18期141-143,共3页
Modern Electronics Technique
关键词
MATLAB
函数逼近
插值
拟合
神经网络
Matlab
function approximation
interpolation
fitting
neural network
