论文部分内容阅读
以初选２２项土壤理化性质作为变量进行主成分分析，经降维后，确定能反映本区土壤分类特征的硅铝铁率比（Ａ１层／Ｂ层）、Ａ１层ＣＥＣ／粘粒比、Ａ１层Ａｌ２Ｏ３（％）、Ｂ层盐基饱和度（％），Ｂ层硅铝铁率和Ａ１层硅铝率等６个主要分类指标，再以主成分分析作初始分类，判别分析确定最终分类。同系统分类比较研究表明，虽然分类原则不同，但指标的内在联系大致吻合。数值分类可为系统分类的诊断特征确定提供一定依据。
The principal component analysis of 22 primary soil physicochemical properties was taken as a variable. After dimension reduction, the ratio of iron to aluminum to iron (A1 / B), CEC / clay ratio of A1, A1 layer Al2O3 (%), B base salt saturation (%), B-layer rate of silicon and aluminum and A1 layer Si-Al rate six primary classification indicators, and then the principal component analysis for the initial classification, discriminant analysis to determine the final classification . A comparative study on the same system shows that although the classification principles are different, the internal relations of the indicators are roughly in agreement. Numerical classification can provide a certain basis for the determination of the diagnostic features of the system classification.