The Application of Visible and Near-Infrared Spectroscopy Combined with Chemometrics in Classification of Dried Herbs

The fast differentiation and classification of herb samples are complicated processes due to the presence many various chemical compounds. Traditionally, separation techniques have been employed for identification quantification compounds present in different plant matrices, but they tedious, time-consuming destructive. Thus, a non-targeted approach would be specifically advantageous this purpose. In study, spectroscopy visible near-infrared range pattern recognition techniques, including principal component analysis (PCA), linear discriminant (LDA), quadratic (QDA), regularized (RDA), super k-nearest neighbor (SKNN) support vector machine (SVM) were applied develop models that enabled discrimination commercial dried herbs, mint, linden, nettle, sage chamomile. error rates validation data below 10% all methods, except SKNN. results obtained confirm methods constitute good non-destructive tool rapid species can used routine quality control by pharmaceutical industry, as well herbal suppliers, avoid mislabeling.