摘要
本文对工程中经常遇到的非线性最小二乘拟合问题提出了一个计算机算法，该方法共分两步：首先利用优选法求出非线性参数，然后计算线性参数。对于三参数曲线，利用线性回归分析求线性参数；对于多参数曲线，则利用广义逆矩阵理论求出线性参数。在这些算法中，迭代初值可以任意选取，而迭代总是收敛的。
A computational method on nonlinear least square fit problem,which is often en-countered in engineering,is presented in this paper.The method contains two steps:at firstthe nonlinear parameters are found by the optirnization method,and then linear parameterscan be achieved by linear regression analysis for the curve with three parameters and by thetheory of generalized inverse matrix for the curve with more than three parameters. For thesealgorithms,the initial values for iterative computations can be selected arbitrarily and the it-eration is always convergent.
出处
《中国铁道科学》
EI
CAS
CSCD
北大核心
1995年第3期64-71,共8页
China Railway Science
关键词
非线性拟合
线性回归
工程数学
最小二乘法
Nonlinear least square fit,Optimization method,Linear regression,generalized inverse matrix
