论文部分内容阅读
测量值的数据处理有重要意义。恒定量的多次重复测量和剔除异常值,可以降低随机误差,提高精度。在二变量测量中,由于测量误差影响,描述随自变量致变的因变量值,并非一条光滑的函数曲线,而是波动或折线形状。运用有关误差理论可以求得表示二变量相互关系的光滑曲线,这种测量值的数学处理方法,称为测量值的拟合。它可以部分地消除测量值中包含的随机误差,提高测量精度。本文说明了拟合的最小二乘法原理,给出了变量测量适用的五点三次拟合方法和计算机源程序;引出了迭代拟合以降低测量值对真值的均方根误差的概念,并比较了五点三次拟合和直线拟合的效果。
The measurement of the data processing is of great significance. A constant amount of repeated measurements and remove abnormal values, can reduce the random error, improve accuracy. In two-variable measurement, due to measurement error, it describes the dependent variable that is dependent on the independent variable. It is not a smooth function curve but a fluctuating or polyline shape. Using the error theory, a smooth curve representing the interrelationship of two variables can be obtained. The mathematical treatment of such measurements is called the fitting of the measured values. It can partially eliminate the random errors contained in the measured values and improve the measurement accuracy. In this paper, we introduce the principle of least-square fitting and give a fitting method of five-point and three-point fitting for variable measurement. The paper introduces the concept of iterative fitting to reduce the root-mean- And compared the five-thirds fit and straight-line fitting effect.