摘要
由于函数y=y0+Aexp(Bx)是非线性的,因此其中参数y0,A,B的拟合较线性回归要复杂得多,并且难以唯一确定,因为它们往往与所给的初始值有关。本文提出了黄金分割与线性回归相结合的迭代方法及其算法。该算法能确保参数的唯一确定,而与拟合时所给的初始值无关。通过对两个常见实例的计算已证明了这一点,并且表明该算法简单、快速而有效。
Compared with linear regression,it is difficult to fit the parameters yo, A,B in the function y=y0+ Aexp(Bx) because of its non-linearity. This paper presents a new method in which gold-section is combined with linear-regression, and the corresponding algorithm. It has been shown, by application to two common examples,that the new fitting method is simple,quick and effective,and results in unique determination of the p.arameters, being independent of the given initial value during fitting.
关键词
黄金分割
线性回归
参数拟合
背底内耗
G-L图
gold-section, linear regression, parameter fitting, background internal friction
G-L plot