论文部分内容阅读
5.多元线性相关分析多元线性相关分析涉及的内容较多,现只着重介绍两方面的问题:一、如何根据几个变量的联合来预测另一变量,以改善预测的精度;二、如阿分析变量之间错综复杂的关系.上述任务,主要通过计算回归方程、复相关系数及偏相关系数来完成.下面以某铂矿氧化带内一探槽中18个样品的Pt、As、Ag、Cu的分析结果为例,说明具体方法.整个计算可分三个阶段:(1)计算相关矩阵所距相关矩阵,所谓相关矩阵,即将变量所有可能的两两组合的相关系数列成表格,如表7.例如Pt-Ag的相关系数
5. Multivariate linear correlation analysis Multivariate linear correlation analysis involves more content, and now focuses on two aspects: First, how to predict another variable based on the combination of several variables to improve the prediction accuracy; Second, Analysis of the complex relationship between the variables.The above tasks, mainly through the calculation of regression equations, complex correlation coefficient and partial correlation coefficient to complete the following to a platinum ore oxidation zone of 18 samples of the trough Pt, As, Ag, Cu (1) Calculate the correlation matrix from the correlation matrix, the so-called correlation matrix, the variables will be a combination of all possible combinations of two or two as a table, as shown in Table For example, the correlation coefficient of Pt-Ag