Nardostachys jatamansi DC., a herb endemic to the Himalayas, has a long history of use in ayurvedic and other traditional medicine systems. The quality of N. jatamansi materials is closely related to their geographical origins. This study aims to develop a method for quickly and accurately identifying the geographical origin of N. jatamansi samples. A total of 118 samples of N. jatamansi samples were collected from five production areas in China. High performance liquid chromatography (HPLC), inductively coupled plasma mass spectrometry (ICP-MS) and gas chromatography-mass spectrometry (GC–MS) were used to characterize the active compounds, mineral elements, and volatile components in N. jatamansi, and four algorithms were introduced to explore the potential of various machine learning modeling techniques for geographical origin authentication. Three active compounds, 8 mineral elements and 24 volatile metabolites in N. jatamansi were identified to be the key information for origin traceability and authentication. The fused data based on the 3 active compounds and the other feature variables led to the highest accuracies for origin classification, with the accuracies of linear discriminant analysis (LDA), K-nearest neighbors (KNN), support vector machine (SVM), and random forest (RF) are 94.44 %, 100 %, 100 %, and 100 %, respectively. This study demonstrates the effectiveness of a multi-source data fusion strategy for geographical origin traceability of N. jatamansi.
Read full abstract