基于Matlab上实现函数逼近.pdfVIP

2009 1 305 基于Matlab 实现函数逼近 曾德惠 ( 445000) : 为满足工程应用中对数据处理的需要, 讨论基于Mat lab 实现函数逼近的三种方 : 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 的Matlab 实现方 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 , 利用Mat lab 编程可以简单有效地实现函数逼近 : Mat lab; 函数逼近; 插值; 拟合; 神经网络 :T P391. 9; O241 :A : 1004-373X( 2009) 1- 14 1-03 Realization of Function Approximation Based on Matlab ZENG Dehui (School of Science, Hubei Institut e for Nat ionalit ies, Enshi,445000, China) Abstract :Three means to realize function approach such as the interpolation approach,f itting approach as w ell as the neu- ral netw ork approach are discussed based on Matlab t o meet the demand of data processing in engineering application. Based on basic principle of the introduction, realization methods to piecew ise linear interpolation, H ermite interpolation, cubic spline in- terpolat ion and least squares curve fitting method are researched using interpolation funct ion and fitting function in Matlab with example. BP neural network to proximate nonlinear functions is designed, desired efect is achieved through the training and simulation of net work.All results indicate that different methods are adopted according to specific problems, and function approach is realized simply and effectively in Matlab. Keyw ords: Matlab; function approximat ion; interpolation;fitting; neural network 0 1 , , , , , , p (x ) , f ( x ) , ,