粗糙集聚类分析对姜黄属植物数值分类的研究Ι

Numerical Taxonomy of Medical Plant of Curcuma L. with Rough-set Theory

以姜科姜黄属(CurcumaL.)的31种植物为研究对象,把47项分类性状(指标)分解为定性指标和定量指标,将定性指标进一步分解为二元性状指标和三元性状指标,然后各自进行粗糙集理论的处理和聚类,把聚类结果按交集分类的办法形成定性指标综合分类的结果,分出8大类,并与肖小河等用47项指标进行主成分析和聚类分析的结果加以比较。结果表明:把分类性状(指标)分解是合理的、正确的,所分出的结果较为精确,与实际更为相符;粗糙集理论与聚类分析相结合是非常有意义的。

31 medical plants of Curcuma L. were studied in this experiment. Firstly, 47 taxonomy character were decomposed m quahtative index and quantitative index, Secondly, the qualitative character were decomposed in two-state character and three-state character. And thirdly, the rough-set theory and cluster synthesis were used to treat these data, respectively and the result was to form 8 kinds of qualitative index synthesis taxonomy, by intersection taxonomy. The results were compared with those of Xiaoxiao-he made in the principal component and cluster analysis.