论文部分内容阅读
中药色谱指纹图谱具有成分复杂、图谱基线和成分保留时间易受外界因素干扰等特点。本文对获得的色谱指纹图谱进行基线校正和色谱峰保留时间校正的数据预处理,不需色谱峰的积分,直接用各色谱信号的全谱数据评价相似度、主成分投影分析和谱系聚类分析,将各厂家的注射液样本正确分类并检出其中的异常样本,为处理和评价中药色谱指纹图谱数据,提供客观、便捷的方法。
Chromatographic fingerprints of traditional Chinese medicine are complex in composition, and the baseline of the map and the retention time of the components are easily disturbed by external factors. In this paper, baseline chromatograms and chromatographic peak retention time corrections are performed on the obtained chromatographic fingerprints. No chromatographic peaks are integrated. The whole spectrum data of each chromatogram signal are used to evaluate similarity, principal component projection analysis and pedigree cluster analysis. The correct classification of injection samples from various manufacturers and the detection of abnormal samples among them provide an objective and convenient method for processing and evaluating the chromatographic fingerprint data of traditional Chinese medicine.