论文部分内容阅读
采用一种新型模式识别方法——双指标等级序列模式识别法,对不同厂家生产的中药复方十全大补丸样品氯仿提取物的红外指纹图谱进行了分析,以实现中药复方的准确鉴别及质量控制。该法同时具有无监督及有监督模式识别的优点,无须经验知识,通过确定每个样品双指标序列中的相似样品序列,同时进行聚类及分类分析。采用该法对3个不同厂家生产的22个中药复方中成药十全大补丸样品的氯仿提取物的红外指纹图谱进行鉴别研究,正确识别率达100%,而采用UPGMA系统聚类方法的正确识别率仅为72.7%。双指标等级序列模式识别法结合红外指纹图谱,无需任何先验知识即可实现中药复方的快速准确鉴别及质量评价。
A novel pattern recognition method, the double-index hierarchical sequential pattern recognition method, was used to analyze the infrared fingerprints of the chloroform extracts of traditional Chinese medicine compound Shiquandabu pills produced by different manufacturers, in order to realize the accurate identification and quality of traditional Chinese medicine compound formulas. control. This method has the advantages of unsupervised and supervised pattern recognition at the same time. It does not need experience and knowledge to determine the similar sample sequence in the double-indicator sequence of each sample and perform clustering and classification analysis at the same time. This method was used to identify the infrared fingerprints of the chloroform extracts of 22 Chinese herbal compound Shiquandabu Pill samples from 3 different manufacturers. The correct recognition rate was 100%, and the UPGMA system clustering method was used correctly. The recognition rate is only 72.7%. The dual-index hierarchical sequence pattern recognition combined with infrared fingerprinting can realize rapid and accurate identification and quality evaluation of Chinese herbal compound without any prior knowledge.