Skip to main content
Fig. 1 | Chinese Medicine

Fig. 1

From: Machine learning integration of multi-modal analytical data for distinguishing abnormal botanical drugs and its application in Guhong injection

Fig. 1

HPLC–UV data analysis. A Comparison of aceglutamide (peak 2) and hydroxysafflor yellow A (peak 16) between normal (N25) and abnormal (A1) batches, the data showed that the peak shape of peak 2 and the absolute integral values of peak 16 were significantly changed. B Peak 16 was used as the reference peak and relative peak area (RPA) was calculated as the value of the AUC of the target peak (At) dividing by the AUC of the reference peak (As). C Pearson correlation coefficient between batches was analyzed by involving the RPA values calculated. It showed that only sample A20 was recognized as the abnormal sample by setting the threshold value at 0.9. D PLS-DA using the RPA values: nine normal samples (in the dashed ellipse) were out of the cluster (green area) along with one normal (N31) and two abnormal (A24 and A26) samples falling outside of the 95% confidence interval

Back to article page