Digital Identification of Ginseng Growth Years Based on Support Vector Machine and UPLC-QTOF-MS
Objective:To conduct digital identification and analysis of ginseng growth years based on ultra-performance liquid chromatography-quadrupole time-of-flight mass spectrometry(UPLC-QTOF-MS)analysis combined with support vector machine(SVM)for data modeling.Methods:Ginseng samples aged 3,4,5,and 15 years were analyzed using UPLC-QTOF-MS.Peak correction,extraction,and quantization were performed using mixed quality control samples as a reference,obtaining exact mass-retention time(EMRT)data pairs that reflected chemical composition information.The data were then modeled using SVM.Model evaluation was carried out based on 5-fold,10-fold,and 20-fold internal cross-validation using parameters such as accuracy(Acc),precision(P),and area under the curve(AUC).The growth years of ginseng were identified based on the established data model.Results:After quantization,6556 EMRTs were obtained from 80 batches of ginseng.The data model built using SVM exhibited excellent identification performance,with Acc,P,and AUC all exceeding 0.900,and the external identification verification accuracy was 100%.Conclusion:UPLC-QTOF-MS analysis combined with the SVM algorithm can efficiently and accurately achieve digital identification of ginseng growth years,providing a reference for exploring the identification of the growth years of Chinese herbal medicines and contributing to their quality control.
国家科技期刊平台
NSTL
万方数据
