论文部分内容阅读
本文应用聚类判别分析的多元统计方法,对十二个厂家生产的安宫牛黄丸进行了综合质量评价。通过气相色谱法、紫外-可见二阶导数光谱法、薄层扫描法、显微镜检等方法获得判别指标,以Bayes准则的逐步判别分析方法将诸厂家生产的安宫牛黄丸分为优、中、劣三级。判别结果较为切合实际。实践证明,在多指标的情况下,利用此方法,可以克服经验鉴别的主观性和片面性,比较客观地反映中成药的内在质量。
This article applies the multivariate statistical method of cluster analysis to evaluate the comprehensive quality of Angong Niuhuang Pills produced by twelve manufacturers. Discriminant indices were obtained by gas chromatography, UV-visible second derivative spectroscopy, thin-layer scanning, microscopy, and other methods. The Bayesian criteria for stepwise discriminant analysis were used to classify Angong Niuhuang Pills produced by various manufacturers into superior, middle, and Bad three. The discrimination result is more realistic. Practice has proved that in the case of multiple indicators, the use of this method can overcome the subjectivity and one-sidedness of empirical identification and more objectively reflect the inherent quality of proprietary Chinese medicines.