Rapid identification and quantification of adulteration in Dendrobium officinale using nuclear magnetic resonance spectroscopy combined with least-squares support vector machine

Dendrobium officinale (D. officinale) is commonly used as a functional food or herbal medicine worldwide, but often adulterated with low-priced materials. In this study, we attempted to develop an integrated method of least-squares support vector machine and nuclear magnetic resonance spectroscopy (LS-SVM-NMR) to identify and quantify the adulteration of D. officinale powder. We found that LS-SVM-NMR can identify the adulterated D. officinale powder with an overall accuracy of 100% and quantify its purity with an R-2 of 0.999 and RMSE of 1.410 at model validation phase. In addition, our results from the double-blinded test revealed that LS-SVM-NMR can yield a classification accuracy of 98% for distinguishing the adulterated D. officinale powder, and predict its purity with an R-2 of 0.998 and RMSE of 2.660. Moreover, adulteration evaluation reports can be obtained in 10 min. Therefore, LS-SVM-NMR proposed herein could be used as a promising and time-saving tool for the quality evaluation of D. officinale or even other food products.