多元数量区划法与应用

主坐标区划法是将聚类分析与主成分分析方法结合起来,形成一种独特的标定方法,进行数量分类(区)。非线性映象区划法,是通过保持原始数据结构进行降维的方式。在限定由高维空间到低维空间的变化保持连续性的前提下,通过迭代运算保持数据的几何结构相对不变,最终达到降维投影的目的。本文以辽宁省凤城县和沈阳地区农业气候区划为例,说明了这两种多元数量分类区划法在农业中的实际应用。

Multivariate Quantitative Classification and Its Application

Principal coordinate clustering, which combines clustering analysis with principal component analysis, is a specific algorithm used in quanti- tative classification, while nonlinear mapping is an algorithm that lowers the number of dimensions of source data by stepwise calculations in which the dimension lowering procedure must be continuous and the geometrical structure of data remains unchanged. The paper uses the agrometeorological zonings of Fengcheng County and Shenyang region of Liaoning Province as examples to illustrate the practi- cal application of the two multivariate quantitative classification methods in agriculture.