作者: Qiang Zhang , Liang-Liang Zhang , Jian-Guo Xu , Guo-Ting Cui
DOI: 10.1007/S11694-019-00257-7
关键词:
摘要: In this paper, image analysis and electronic nose were used to develop a rapid, reliable undamaged method of identifying the Semen cuscutae its adulterants including radish seed Sinapis alba seeds. The results showed that highest identification rate was 100% for training set 96.5% test based on various chemometric techniques principal component analysis, linear discriminant (LDA), k-nearest neighbor, random forests, artificial neural network support vector machine analysis. LDA exhibited better discrimination result, ranging from 95.5–100% correct classification 95.4–100% cross-validation rate, respectively. model data 16 30 s best, both reached 100%. These provided simple, fast non-destructive method identify true false cuscutae, which can serve as reference identify the authenticity medicinal plants.