论文部分内容阅读
作者采用R-型因子分析(主成分分析)及R-型聚类分析对芳环上166种取代基的六种化学结构参数进行计算。仅由参数的相关系数矩阵及R-型聚类分析的计算结果均不足以充分说明此等参数的相互独立性,但通过R-型因子分析采用方差极大正交旋转及Promax斜旋转则能判定该6种参数的相互独立性,为此等参数的选用提供了确切的论据。作者比较了R-型因子得分的取代基分群与Q-型聚类分析的分群结果。指出两种分群方法各有其重要作用。
The authors used R-type factor analysis (principal component analysis) and R-type cluster analysis to calculate the six chemical structure parameters of 166 substituents on aromatic rings. Only by the correlation coefficient matrix of parameters and the results of R-type cluster analysis are not enough to fully explain the mutual independence of these parameters, but by R-factor analysis using orthogonal variational rotation and Promax oblique rotation can Determine the mutual independence of the six kinds of parameters, for the selection of these parameters provide an exact argument. The authors compared the clustering results for the R-type factor scores with the Q-type cluster analysis. It points out that each of the two clustering methods has its own important role.