[1]张建华..关于相关系数作拟合优度判据的研究[J].郑州大学学报(工学版),2001,22(03):72-74.[doi:10.3969/j.issn.1671-6833.2001.03.022]
　ZHANG Jianhua..A study on the correlation coefficient as a goodness-of-fit criterion[J].Journal of Zhengzhou University (Engineering Science),2001,22(03):72-74.[doi:10.3969/j.issn.1671-6833.2001.03.022]

点击复制

关于相关系数作拟合优度判据的研究()

分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:

22卷

期数:

2001年03期

页码:

72-74

栏目:

出版日期:

1900-01-01

文章信息/Info

Title:

A study on the correlation coefficient as a goodness-of-fit criterion

作者:

张建华.

郑州大学工程力学系,

Author(s):

ZHANG Jianhua.

关键词:

相关系数;  拟合;  优度;  回归

Keywords:

correlation coefficient;  fitting;  Superiority;  Regression

DOI:

10.3969/j.issn.1671-6833.2001.03.022

文献标志码:

A

摘要:

实验数据处理,常用相关系数的大小来表示线性拟合优劣的判据.通过实验例证分析研究,相关系数好不一定直线拟合结果就好,实验结果线性拟合得好不一定相关系数就很大,说明不宜用相关系数作实验结果拟合优度的判据.只有它为回归系数时,或当有一变量不可能用一个固定的标度来测量时,才有实际意义的.

Abstract:

In experimental data processing, the magnitude of the correlation coefficient is often used to express the criterion of linear fit. Through experimental illustration analysis and research, a good correlation coefficient does not necessarily mean a good straight-line fitting result, and a good linear fitting of experimental results does not necessarily mean a large correlation coefficient, indicating that the correlation coefficient should not be used as a criterion for the goodness-of-fit of experimental results. It is only meaningful if it is a regression coefficient, or when there is a variable that cannot be measured on a fixed scale.

常用功能