论文部分内容阅读
基于主灵敏度矢量稳健主成分回归(RPPSV)和蒙特卡罗交互检验(MCCV)方法对29个不同来源的银杏提取物和33 个不同产地的灯盏花提取物的色谱指纹图谱进行质量评价,分别定性判断出银杏样本数据集中的7个异常样本和灯盏花样本数据集中的4个异常样本。与相关系数法相比,RPPSV具有更好的异常样本检测能力,但是两者评价指纹图谱的结果基本是一致的。
Based on the main sensitivity vector robust principal component regression (RPPSV) and Monte Carlo interaction test (MCCV) method, the chromatographic fingerprints of 29 Ginkgo biloba extracts and 33 Erigeron breviscapus extracts from different origins were evaluated for quality. Seven abnormal samples in the Ginkgo sample data set and four abnormal samples in the Erigeron sample data set were determined. Compared with the correlation coefficient method, RPPSV has better abnormal sample detection ability, but the results of the two evaluation fingerprints are basically the same.