一、引言一般而论,一元回归法往往相关系数值较小,精度较差,而逐步回归法则计算比较烦琐,工作量较大。因此,如何建立一个计算简单,而相关性又较好的回归方程,是一个值得探讨的问题。笔者在对柘林主坝沉陷资料进行回归分析时,发现对某些常见的函数曲线,适当增加一个待定系数C,可以通过计算机的运算技巧而得到一个相关系数最大的非线性一元回归方程。 I. INTRODUCTION In general, the univariate regression method tends to have small correlation coefficient values ​​and poor accuracy, while the step-by-step regression rule is more cumbersome and requires a lot of work load. Therefore, how to set up a regression equation with simple calculation and good relativity is a problem worth exploring. When regression analysis of subsidence data of Zhelin main dam, the author found that for some common function curves, adding a coefficient C to be determined appropriately, a nonlinear regression equation with the largest correlation coefficient can be obtained through calculation skills of the computer.