中文摘要：

本文首次将集对原理用于天然草地分类中，以中国新疆某地区的31个天然草地为例进行计算，并将分析结果，与灰色关联度分析法、投影寻踪法以及投影寻踪动态聚类法作了对比分析。结果表明，集对法与三种方法分析结果一致，但集对法计算简单，使用方便，不受样本容量限制，且能清晰和形象地显示所属类别的程度，在多因素聚类分析中具有广阔的应用前景。

英文摘要：

The principle of set pair analysis （SPA） was attempted to apply to the classification of natural grassland quality firstly in this paper. The data from 31 natural grasslands in Sinkiang China were used for the calculation. The result was compared with that from the Grey Relational Analysis, Projection Pursuit method and Projection Pursuit Dynamic Cluster method （PPDC）. It showed that their outcomes were consistent. While compared with the latter three methods, the method of Set Pair Analysis was simple, convenient and not restricted of sample size. Especially it clearly and vividly displayed the relational degree to a class. Therefore it has a wide prospect of application in multi-factor cluster analysis.