Geographical origins of Angelica sinensis using functional compounds and multielement with machine learning-based...

Ensuring food traceability is essential for maintaining safety and authenticity. Angelica sinensis (Oliv.) Diels (AS), a medicinal food prized for its rich nutritional value and tonic effects, is frequently vulnerable to geographic origin fraud. In this study, 16 functional compounds and 40 multielement were utilized to investigate the regional characteristics and the geographical origins authentication of AS samples from 8 different origins. Three algorithms were introduced, and the K-nearest neighbors (KNN) model constructed by the second-level fusion using 22 key variables screened by VIP features performed the best for AS origin classification, with a prediction accuracy of 100.00 % in both the training set and the testing set. Moreover, 5 environmental factors, including longitude, latitude, cation exchange capacity, accumulated temperature of growing degree days above 5 °C and dry or moisture index, were identified as the primary influenced environmental factors.