论文部分内容阅读
目的 :探讨用聚类判别分析方法评价六味地黄丸质量。方法 :对六味地黄丸及其模拟样品进行薄层扫描分析 ,将所得特征数据利用聚类分析方法进行分类 ,用逐步判别分析方法优选特征数据、建立判别函数 ,并对分类结果进行检验。结果 :样品及模拟品被分为 4类 ,分类合理 ,判别函数回判准确率为 10 0 %。结论 :聚类判别分析方法可较好地评价六味地黄丸的质量
Objective: To evaluate the quality of Liuwei Dihuang Wan by cluster analysis. Methods: The thin layer scanning analysis was performed on Liuwei Dihuang Pills and its simulated samples. The obtained characteristic data were classified using the cluster analysis method. Stepwise discriminant analysis method was used to select the characteristic data, discriminate function was established, and the classification results were tested. Results : Samples and simulated products were divided into 4 categories. The classification was reasonable, and the discriminant function accuracy rate was 100%. Conclusion : Cluster analysis method can evaluate the quality of Liuwei Dihuang Wan