论文部分内容阅读
地学研究经常涉及两变量间的线性回归分析。此类线性关系可有三种情形：（１）ｘ数据的观测误差很小，数据点的离散只与ｙ数据有关；（２）ｘ和ｙ数据都有观测误差，且数据点的离散仅由观测误差所引起；（３）数据点的离散不仅与观测误差有关，而且还受到其它因素的影响（但两变量仍存在显著的线性关系）。其中第三种情形在地学中最为常见，但已有的回归分析方法只是针对前两种情形。本文对ＬＮＳ方法进行了改进，使之适用于第三型的回归分析。改进的方法在逻辑上具有一致性，在应用上适合线性关系第三型线性关系的特点，因而优于传统的ＯＬＳ方法
Geoscience research often involves linear regression analysis between two variables. There are three cases of such linear relationship: (1) the observation error of x data is very small, and the dispersion of data points is only related to y data; (2) Both x and y data have observation errors, and the data points are separated only by observation (3) Discrete data points are not only related to observation errors but also influenced by other factors (but there is still a significant linear relationship between the two variables). The third case is the most common in geography, but the existing regression analysis method is only for the first two cases. In this paper, the LNS method has been improved, making it suitable for the third type of regression analysis. The improved method is logically consistent and adapts linearly to the third linear relation in the application, which is superior to the traditional OLS method